Package 'chillR'

Description

chillR contains functions for processing temperature records into chilling (Chilling Hours, Utah Chill Units and Chill Portions) and heat units (Growing Degree Hours). Regarding chilling metrics, Chill Portions are often considered the most promising, but they are difficult to calculate. This package makes it easy. chillR also contains procedures for conducting a Partial Least Square regression analysis relating phenological dates (e.g. bloom dates) to either mean temperatures or mean chill and heat accumulation rates, based on long-term weather and phenology records (Luedeling and Gassner 2012). As of version 0.65, it also includes functions for generating weather scenarios with a weather generator, for conducting climate change analyses for temperature-based climatic metrics and for plotting results from such analyses. Since version 0.70, chillR contains a function for interpolating hourly temperature records. As of version 0.72.6, the package contains functions to design elegant plots using the ggplot2 package.

Author(s)

References

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52. doi:10.1016/j.agrformet.2011.10.020

Description

Takes the ‘Year', 'Month', 'Day' and, if available, 'Hour', 'Minute' and 'Second' columns of a data.frame and uses them to produce a 'Date' column that uses R’s standard data/time format.

Usage

add_date(df)

Arguments

Details

Converts YEARMODA to R date

Value

data.frame consisting of the df input and a new column 'Date'

Author(s)

Eike Luedeling

Examples

add_date(KA_weather)
add_date(Winters_hours_gaps)

Description

This is a pretty rudimentary function to predict phenological dates from chilling and forcing requirements and hourly chilling and forcing data. Note that there are enormous uncertainties in these predictions, which are hardly ever acknowledged. So please use this function with caution.

Usage

bloom_prediction(
  HourChillTable,
  Chill_req,
  Heat_req,
  Chill_model = "Chill_Portions",
  Heat_model = "GDH",
  Start_JDay = 305
)

Arguments

Details

This function is a bit preliminary at the moment. It will hopefully be refined later.

Chill metrics are calculated as given in the references below. Chilling Hours are all hours with temperatures between 0 and 7.2 degrees C. Units of the Utah Model are calculated as suggested by Richardson et al. (1974) (different weights for different temperature ranges, and negation of chilling by warm temperatures). Chill Portions are calculated according to Fishman et al. (1987a,b). More honestly, they are calculated according to an Excel sheet produced by Amnon Erez and colleagues, which converts the complex equations in the Fishman papers into relatively simple Excel functions. These were translated into R. References to papers that include the full functions are given below. Growing Degree Hours are calculated according to Anderson et al. (1986), using the default values they suggest.

Value

data frame containing the predicted dates of chilling requirement fulfillment and timing of the phenological stage. Columns are Creqfull, Creq_year, Crey_month, Creq_day and Creq_JDay (the row number, date and Julian date of chilling requirement fulfillement), Hreqfull, Hreq_year, Hreq_month, Hreq_day and Hreq_JDay (the row number, date and Julian date of heat requirement fulfillment - this corresponds to the timing of the phenological event.

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours or Utah Models, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Model references:

Chilling Hours:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Utah Model:

Richardson EA, Seeley SD, Walker DR (1974) A model for estimating the completion of rest for Redhaven and Elberta peach trees. HortScience 9(4), 331-332

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Model comparisons and model equations:

Luedeling E, Zhang M, Luedeling V and Girvetz EH, 2009. Sensitivity of winter chill models for fruit and nut trees to climatic changes expected in California's Central Valley. Agriculture, Ecosystems and Environment 133, 23-31

Luedeling E, Zhang M, McGranahan G and Leslie C, 2009. Validation of winter chill models using historic records of walnut phenology. Agricultural and Forest Meteorology 149, 1854-1864

Luedeling E and Brown PH, 2011. A global analysis of the comparability of winter chill models for fruit and nut trees. International Journal of Biometeorology 55, 411-421

Luedeling E, Kunz A and Blanke M, 2011. Mehr Chilling fuer Obstbaeume in waermeren Wintern? (More winter chill for fruit trees in warmer winters?). Erwerbs-Obstbau 53, 145-155

Review on chilling models in a climate change context:

Luedeling E, 2012. Climate change impacts on winter chill for temperate fruit and nut production: a review. Scientia Horticulturae 144, 218-229

The PLS method is described here:

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52.

Wold S (1995) PLS for multivariate linear modeling. In: van der Waterbeemd H (ed) Chemometric methods in molecular design: methods and principles in medicinal chemistry, vol 2. Chemie, Weinheim, pp 195-218.

Wold S, Sjostrom M, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab 58(2), 109-130.

Mevik B-H, Wehrens R, Liland KH (2011) PLS: Partial Least Squares and Principal Component Regression. R package version 2.3-0. http://CRAN.R-project.org/package0pls.

Some applications of the PLS procedure:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Yu H, Luedeling E and Xu J, 2010. Stronger winter than spring warming delays spring phenology on the Tibetan Plateau. Proceedings of the National Academy of Sciences (PNAS) 107 (51), 22151-22156.

Yu H, Xu J, Okuto E and Luedeling E, 2012. Seasonal Response of Grasslands to Climate Change on the Tibetan Plateau. PLoS ONE 7(11), e49230.

The exact procedure was used here:

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

hourtemps <- stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year > 2008), ]),
                                latitude=50.4)

CT <- chilling_hourtable(hourtemps, Start_JDay = 305)

bloom_prediction(CT, Chill_req = 60, Heat_req = 5000, Chill_model = "Chill_Portions",
                 Heat_model = "GDH", Start_JDay = 305)

Description

This is a pretty rudimentary function to predict phenological dates from chilling and forcing requirements and hourly chilling and forcing data. Note that there are enormous uncertainties in these predictions, which are hardly ever acknowledged. So please use this function with caution.

Usage

bloom_prediction2(
  HourChillTable,
  Chill_req,
  Heat_req,
  permutations = FALSE,
  Chill_model = "Chill_Portions",
  Heat_model = "GDH",
  Start_JDay = 305,
  infocol = NULL
)

Arguments

Details

This function is an update to the bloom_prediction function, which was quite slow and didn't allow testing multiple chilling and heat requirements. In this updated version, chilling and heat requirements can be supplied as vectors, which are interpreted in sequence, with each pair of Chill_req and Heat_req values matched according to their position in the vectors. Through the permutations argument, it is also possible to compute stage occurrence dates for all possible combinations of the requirements specified by the Chill_req and Heat_req vectors.

The model allows specifying any numeric column as the chill and heat columns, indicated by the Chill_model and Heat_model parameters.

Value

data frame containing the predicted Julian dates of chilling requirement fulfillment and timing of the phenological stage. Columns are Season, Creq, Hreq, Creq_full (day when the chilling requirement is fulfilled) and Pheno_date (the predicted date of the phenological event).

Author(s)

Eike Luedeling

References

Model references:

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Examples

hourtemps <- stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year > 2008), ]),
                                latitude = 50.4)

CT <- chilling_hourtable(hourtemps, Start_JDay = 305)

bloom_prediction2(CT, c(30, 40, 50), c(1000, 1500, 2000))
bloom_prediction2(CT, c(30, 40, 50), c(1000, 1500, 2000), permutations = TRUE)

Description

This is a pretty rudimentary function to predict phenological dates from chilling and forcing requirements and hourly chilling and forcing data. Note that there are enormous uncertainties in these predictions, which are hardly ever acknowledged. So please use this function with caution.

Usage

bloom_prediction3(
  hourtemps,
  Chill_req,
  Heat_req,
  models = c(Chill_Portions = Dynamic_Model, GDH = GDH_model),
  permutations = FALSE,
  Chill_model = "Chill_Portions",
  Heat_model = "GDH",
  Start_JDay = 305,
  infocol = NULL
)

Arguments

Details

This function is an update to the bloom_prediction and bloom_prediction2 functions. This version takes hourly temperatures as input rather than requiring pre-calculated chill and heat records. This functionality is now integrated in the function, so that users can now specify a list of temperature metrics/models to be computed and used in the bloom prediction.

Value

data frame containing the predicted Julian dates of chilling requirement fulfillment and timing of the phenological stage. Columns are Season, Creq, Hreq, Creq_full (day when the chilling requirement is fulfilled) and Pheno_date (the predicted date of the phenological event).

Author(s)

Eike Luedeling

References

Model references:

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Examples

hourtemps <- stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year > 2007), ]),
                                latitude = 50.4)

bloom_prediction3(hourtemps, c(30, 140, 50), c(1000, 1500, 2000))

bloom_prediction3(hourtemps, c(30, 40, 50), c(1000, 1500, 2000), permutations = TRUE,
                  Start_JDay = 1)

bloom_prediction3(hourtemps, c(300, 400, 600), c(100, 150, 200), permutations = TRUE,
                  Start_JDay = 1, models = c(CH = Chilling_Hours, Heat = GDD),
                  Chill_model = "CH", Heat_model = "Heat")

Description

This function bootstraps the residuals of a 'phenologyFit'. It internally calls 'phenologyFitter' on each bootstrap replicate.

Usage

bootstrap.phenologyFit(
  object,
  boot.R = 99,
  control = list(smooth = FALSE, verbose = FALSE, maxit = 1000, nb.stop.improvement =
    250),
  lower,
  upper,
  seed = 1766588
)

Arguments

Details

bootstrap an object of S3 class 'phenologyFit'

Value

Invisibly returns a list with elements 'boot.R', 'object', 'seed', 'residuals', 'lower', 'upper', and 'res'. The latter list 'res' has 'boot.R' elements, which are lists again. Each of these lists contains named elements 'par', 'value', 'bloomJDays', and 'pbloomJDays'. 'par' are the best fit parameters on the particular bootstrap replicate, 'value' the corresponding RSS, 'bloomJDays' the re-sampled data and 'pbloomJDays' the predicted bloom JDays for this sample.

Author(s)

Carsten Urbach <[email protected]>

Description

Concatenate bootstrap_phenologyfit objects

Usage

## S3 method for class 'bootstrap_phenologyFit'
c(...)

Arguments

Value

An object of class 'bootstrap_phenologyFit', the concatenation of the list of input object.

Description

This is a list of weather stations in California that are contained in the UC IPM database. This can also be generated with make_california_UCIPM_station_list(), but this takes quite a while. So this dataset is supposed to be a shortcut to this.

Format

a data.frame containing stations from the California UC IPM database (), with the columns: "Name", "Code", "Interval", "Lat", "Long", "Elev".

list("Name")

name of the weather station

list("Code")

code of the weather station, indicating the name and the database it comes from

list("Interval")

period of available data (as character string)

list("Lat")

latitude of the station

list("Long")

longitude of the station

list("Elev")

elevation of the station

Source

UC IPM website: http://www.ipm.ucdavis.edu/WEATHER/index.html

Examples

data(california_stations)

Description

This function performs basic tests to determine whether a temperature record complies with chillR's formatting rules. If desired, the function also checks whether the record is complete (has rows for all time units in the interval) and how many values are missing.

Usage

check_temperature_record(
  weather,
  hourly = FALSE,
  completeness_check = TRUE,
  no_variable_check = FALSE
)

Arguments

Value

list containing the following elements: 'data_frequency' ("daily or "hourly), 'weather_object' (boolean, indicates whether records are in a sub-object called weather), 'chillR_compliant' (boolean, indicates whether the object was found to conform to chillR format standards) and 'error' (contains error messages generated during the checking procedure).

Note

This function doesn't check whether there are faulty data. It only tests whether the data is compatible with the requirements of chillR's major functions.

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

check_temperature_record(KA_weather)

Description

chillR's temperature generation procedures require absolute or relative temperature scenarios. This function checks these scenarios for consistency, regarding the data format, the reference year, and whether they are relative or absolute scenarios (based on specified criteria).

Usage

check_temperature_scenario(
  temperature_scenario,
  n_intervals = 12,
  check_scenario_type = TRUE,
  scenario_check_thresholds = c(-5, 10),
  update_scenario_type = TRUE,
  warn_me = TRUE,
  required_variables = c("Tmin", "Tmax")
)

Arguments

Details

Besides being able to validate classic temperature scenarios consisting of "Tmin" and "Tmax" data, the function can also validate other datasets (e.g. outputs of the getClimateWizardData function). To do this, the required variables should be provided as "required_variables" parameter. If there is no column "GCM" in the data element of the scenario, then the check_scenario_type parameter should be set to FALSE.

Value

temperature scenario object, consisting of the following elements: 'data' = a data frame with n_intervals elements containing the absolute or relative temperature information. 'reference_year' = the year the scenario is representative of. 'scenario_type' = the scenario type ('absolute' or 'relative'); 'labels' = and elements attached to the input temperature_scenario as an element names 'labels'.

The function also returns warnings, where elements are missing or the scenario_type appears to be wrong, and it stops with an error, if the scenario isn't specified in a format that is usable by chillR.

Author(s)

Eike Luedeling

Examples

temperature_scenario <- list(data = data.frame(Tmin = c(-5, -2, 0, 4, 9, 12,
                                                        15, 13, 12, 9, 4, 0),
                                               Tmax = c(0, 4, 8, 12, 15, 18,
                                                        21, 19, 17, 14, 11, 5)),
                             reference_year = 1975,
                             scenario_type = "absolute",
                             labels = list(GCM = "none",
                                           RCM = "none",
                                           Time = "1950-2000"))

checked_temperature_scenario <- 
        check_temperature_scenario(temperature_scenario,
                                   n_intervals = 12,
                                   check_scenario_type = FALSE,
                                   scenario_check_thresholds = c(-5, 10),
                                   update_scenario_type = FALSE)
                                            
checked_temperature_scenario <-
        check_temperature_scenario(temperature_scenario,
                                   n_intervals = 12,
                                   check_scenario_type = TRUE,
                                   scenario_check_thresholds = c(-5, 10),
                                   update_scenario_type = FALSE)
                                            
checked_temperature_scenario <-
        check_temperature_scenario(temperature_scenario,
                                   n_intervals = 12,
                                   check_scenario_type = TRUE,
                                   scenario_check_thresholds = c(-5, 10),
                                   update_scenario_type = TRUE)

Description

RSS to minimise by 'phenologyFitter'

Usage

chifull(par, modelfn, bloomJDays, SeasonList, na_penalty = 365, ...)

Arguments

Details

function to compute the RSS

Description

Convert downloaded weather data into a data frame that makes running other chillR functions easy.

Usage

chile_agromet2chillR(downloaded_weather_file, drop_most = TRUE)

Arguments

Details

Processing the data with this function will make the data work well with the remainder of this package.

Value

a data.frame with weather data, according to the downloaded file provided as input. If drop_most is FALSE, all columns from the original dataset are preserved, although some column names are adjusted to chillR's preferences ("Year","Month","Day","Tmin","Tmax","Tmean","Prec", if these columns are present). If drop_most is TRUE, only columns likely to be of interest to chillR users are retained.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2005),])  # this line is
#only here to make the example run, even without downloading a file

# FOLLOW THE INSTRUCTIONS IN THE LINE BELOW THIS; AND THEN  RUN THE LINE
# AFTER THAT (without the #)
# download an Excel file from the website and save it to disk (path: {X}) 
#weather<-fix_weather(chile_agromet2chillR({x}))

hourtemps<-stack_hourly_temps(weather, latitude=50.4)
chilling(hourtemps,305,60)

Description

Function to calculate three common horticultural chill metrics and one heat metric from stacked hourly temperatures (produced by stack_hourly_temps). Metrics that are calculated are Chilling Hours, Chill Units according to the Utah Model, Chill Portions according to the Dynamic Model and Growing Degree Hours.

Usage

chilling(
  hourtemps = NULL,
  Start_JDay = 1,
  End_JDay = 366,
  THourly = NULL,
  misstolerance = 50
)

Arguments

Details

Chill metrics are calculated as given in the references below. Chilling Hours are all hours with temperatures between 0 and 7.2 degrees C. Units of the Utah Model are calculated as suggested by Richardson et al. (1974) (different weights for different temperature ranges, and negation of chilling by warm temperatures). Chill Portions are calculated according to Fishman et al. (1987a,b). More honestly, they are calculated according to an Excel sheet produced by Amnon Erez and colleagues, which converts the complex equations in the Fishman papers into relatively simple Excel functions. These were translated into R. References to papers that include the full functions are given below. Growing Degree Hours are calculated according to Anderson et al. (1986), using the default values they suggest.

Value

data frame showing chilling and heat totals for the respective periods for all seasons included in the temperature records. Columns are Season, End_year (the year when the period ended), Days (the duration of the period), Chilling_Hours, Utah_Model, Chill_portions and GDH. If the weather input consisted of a list with elements hourtemps and QC, the output also contains columns from QC that indicate the completeness of the weather record that the calculations are based on.

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours or Utah Models, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Model references:

Chilling Hours:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Utah Model:

Richardson EA, Seeley SD, Walker DR (1974) A model for estimating the completion of rest for Redhaven and Elberta peach trees. HortScience 9(4), 331-332

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Model comparisons and model equations:

Luedeling E, Zhang M, Luedeling V and Girvetz EH, 2009. Sensitivity of winter chill models for fruit and nut trees to climatic changes expected in California's Central Valley. Agriculture, Ecosystems and Environment 133, 23-31

Luedeling E, Zhang M, McGranahan G and Leslie C, 2009. Validation of winter chill models using historic records of walnut phenology. Agricultural and Forest Meteorology 149, 1854-1864

Luedeling E and Brown PH, 2011. A global analysis of the comparability of winter chill models for fruit and nut trees. International Journal of Biometeorology 55, 411-421

Luedeling E, Kunz A and Blanke M, 2011. Mehr Chilling fuer Obstbaeume in waermeren Wintern? (More winter chill for fruit trees in warmer winters?). Erwerbs-Obstbau 53, 145-155

Review on chilling models in a climate change context:

Luedeling E, 2012. Climate change impacts on winter chill for temperate fruit and nut production: a review. Scientia Horticulturae 144, 218-229

The PLS method is described here:

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52.

Wold S (1995) PLS for multivariate linear modeling. In: van der Waterbeemd H (ed) Chemometric methods in molecular design: methods and principles in medicinal chemistry, vol 2. Chemie, Weinheim, pp 195-218.

Wold S, Sjostrom M, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab 58(2), 109-130.

Mevik B-H, Wehrens R, Liland KH (2011) PLS: Partial Least Squares and Principal Component Regression. R package version 2.3-0. http://CRAN.R-project.org/package0pls.

Some applications of the PLS procedure:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Yu H, Luedeling E and Xu J, 2010. Stronger winter than spring warming delays spring phenology on the Tibetan Plateau. Proceedings of the National Academy of Sciences (PNAS) 107 (51), 22151-22156.

Yu H, Xu J, Okuto E and Luedeling E, 2012. Seasonal Response of Grasslands to Climate Change on the Tibetan Plateau. PLoS ONE 7(11), e49230.

The exact procedure was used here:

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

# weather <- fix_weather(KA_weather[which(KA_weather$Year > 2006), ])
# hourtemps <- stack_hourly_temps(weather, latitude = 50.4)
# chilling(hourtemps, 305, 60)

chilling(stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year > 2006), ]),
         latitude = 50.4))

Description

This function calculates winter chill for temperate trees according to the Chilling Hours Model.

Usage

Chilling_Hours(HourTemp, summ = TRUE)

Arguments

Details

Chilling Hours are calculated as suggested by Bennett (1949) (all hours with temperatures between 0 and 7.2 degrees C are considered as one Chilling Hour.

Value

Vector of length length(HourTemp) containing the cumulative Chilling Hours over the entire duration of HourTemp.

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Chilling Hours references:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

Chilling_Hours(hourtemps$hourtemps$Temp)

Description

This function calculates cumulative values for three chill metrics and one heat metric for every hour of an hourly temperature record. The count is restarted on a specified date each year.

Usage

chilling_hourtable(hourtemps, Start_JDay)

Arguments

Details

Chill metrics are calculated as given in the references below. Chilling Hours are all hours with temperatures between 0 and 7.2 degrees C. Units of the Utah Model are calculated as suggested by Richardson et al. (1974) (different weights for different temperature ranges, and negation of chilling by warm temperatures). Chill Portions are calculated according to Fishman et al. (1987a,b). More honestly, they are calculated according to an Excel sheet produced by Amnon Erez and colleagues, which converts the complex equations in the Fishman papers into relatively simple Excel functions. These were translated into R. References to papers that include the full functions are given below. Growing Degree Hours are calculated according to Anderson et al. (1986), using the default values they suggest.

Value

data frame consisting of all the columns of the THourly input data frame, plus the following additional columns: Chilling_Hours (cumulative number of Chilling Hours since the last Start_JDay), Chill_Portions (same for units of the Dynamic Models), Chill_Units (same for units of the Utah Model) and GDH (same for Growing Degree Hours).

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours or Utah Models, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Model references:

Chilling Hours:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Utah Model:

Richardson EA, Seeley SD, Walker DR (1974) A model for estimating the completion of rest for Redhaven and Elberta peach trees. HortScience 9(4), 331-332

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Model comparisons and model equations:

Luedeling E, Zhang M, Luedeling V and Girvetz EH, 2009. Sensitivity of winter chill models for fruit and nut trees to climatic changes expected in California's Central Valley. Agriculture, Ecosystems and Environment 133, 23-31

Luedeling E, Zhang M, McGranahan G and Leslie C, 2009. Validation of winter chill models using historic records of walnut phenology. Agricultural and Forest Meteorology 149, 1854-1864

Luedeling E and Brown PH, 2011. A global analysis of the comparability of winter chill models for fruit and nut trees. International Journal of Biometeorology 55, 411-421

Luedeling E, Kunz A and Blanke M, 2011. Mehr Chilling fuer Obstbaeume in waermeren Wintern? (More winter chill for fruit trees in warmer winters?). Erwerbs-Obstbau 53, 145-155

Review on chilling models in a climate change context:

Luedeling E, 2012. Climate change impacts on winter chill for temperate fruit and nut production: a review. Scientia Horticulturae 144, 218-229

The PLS method is described here:

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52.

Wold S (1995) PLS for multivariate linear modeling. In: van der Waterbeemd H (ed) Chemometric methods in molecular design: methods and principles in medicinal chemistry, vol 2. Chemie, Weinheim, pp 195-218.

Wold S, Sjostrom M, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab 58(2), 109-130.

Mevik B-H, Wehrens R, Liland KH (2011) PLS: Partial Least Squares and Principal Component Regression. R package version 2.3-0. http://CRAN.R-project.org/package0pls.

Some applications of the PLS procedure:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Yu H, Luedeling E and Xu J, 2010. Stronger winter than spring warming delays spring phenology on the Tibetan Plateau. Proceedings of the National Academy of Sciences (PNAS) 107 (51), 22151-22156.

Yu H, Xu J, Okuto E and Luedeling E, 2012. Seasonal Response of Grasslands to Climate Change on the Tibetan Plateau. PLoS ONE 7(11), e49230.

The exact procedure was used here:

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2008),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

cht<-chilling_hourtable(hourtemps,20)

Description

chilling and forcing response function for the unified model by Chuine

Usage

ChuineCF(x, a, b, c)

Arguments

Value

Returns a numeric vector.

References

Isabelle Chuine, A Unified Model for Budburst of Trees, J. theor. Biol. (2000) 207

Description

Critical forcing value

Usage

ChuineFstar(Ctot, w, k)

Arguments

Value

Returns a numeric vector.

References

Isabelle Chuine, A Unified Model for Budburst of Trees, J. theor. Biol. (2000) 207

Description

Function to make color schemes for color bar plots in the chillR package. Colors are assigned based on values in two columns of a data frame. One column contains a threshold, below which col3 is assigned. If values are above the threshold, the value in the other column determines the color: col1 if the value is negative, col2 if positive. This function is useful for making the PLS output figures in the chillR package.

Usage

color_bar_maker(column_yn, column_quant, threshold, col1, col2, col3)

Arguments

Value

a vector of colors, which can be used as col argument when making plots

Author(s)

Eike Luedeling

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

PLS_results<-PLS_pheno(
  weather_data=make_all_day_table(KA_weather),
  split_month=6,   #last month in same year
  bio_data=KA_bloom)

colbar<-color_bar_maker(PLS_results$PLS_summary$VIP,PLS_results$PLS_summary$Coef,0.8,
       "RED","DARK GREEN","GREY")

Description

Allows the user to convert a list of change scenarios to a single data.frame or vice versa. If it converts from data.frame to list, the user can decide if the returned list should be flat or structured. In case of a list of change scenarios, the list should have named elements. In case of composite names, the function assumes that the location is the first part of the composite name, composite elements are seperated by dot.

Usage

convert_scen_information(scenario_object, give_structure = TRUE)

Arguments

Value

list / data.frame with relative change scenarios

Author(s)

Lars Caspersen

Examples

## Not run: 
download_cmip6_ecmwfr(scenario = 'ssp1_2_6',
                      area = c(55, 5.5, 47, 15.1),
                      user = 'write user id here',
                      key = 'write key here',
                      model = 'AWI-CM-1-1-MR',
                      frequency = 'monthly',
                      variable = c('Tmin', 'Tmax'),
                      year_start = 2015,
                      year_end = 2100)

download_baseline_cmip6_ecmwfr(
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here',
    key = 'write key here',
    model = 'AWI-CM-1-1-MR',
    frequency = 'monthly',

station <- data.frame(
      station_name = c('Zaragoza', 'Klein-Altendorf', 'Sfax',
      'Cieza', 'Meknes', 'Santomera'),
      longitude = c(-0.88,  6.99, 10.75, -1.41, -5.54, -1.05),
      latitude = c(41.65, 50.61, 34.75, 38.24, 33.88, 38.06))

extracted <- extract_cmip6_data(stations = station)

scenario_df <- gen_rel_change_scenario(extracted)

scenario_list <- convert_scen_information(scenario_df)


## End(Not run)

Description

This function calculates daily chill (with three models) and heat accumulation for every day of an hourly temperature record (best generated with stack_hourly_temps). It includes the option to include calculation of a running mean, which smoothes accumulation curves. Especially for the Dynamic Model, this may be advisable, because it does not accumulate chill smoothly, but rather in steps.

Usage

daily_chill(
  hourtemps = NULL,
  running_mean = 1,
  models = list(Chilling_Hours = Chilling_Hours, Utah_Chill_Units = Utah_Model,
    Chill_Portions = Dynamic_Model, GDH = GDH),
  THourly = NULL
)

Arguments

Details

Temperature metrics are calculated according to the specified models. They are computed based on hourly temperature records and then summed to produce daily chill accumulation rates.

Value

a daily chill object consisting of the following elements

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours or Utah Models, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Model references for the default models:

Chilling Hours:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Utah Model:

Richardson EA, Seeley SD, Walker DR (1974) A model for estimating the completion of rest for Redhaven and Elberta peach trees. HortScience 9(4), 331-332

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Model comparisons and model equations:

Luedeling E, Zhang M, Luedeling V and Girvetz EH, 2009. Sensitivity of winter chill models for fruit and nut trees to climatic changes expected in California's Central Valley. Agriculture, Ecosystems and Environment 133, 23-31

Luedeling E, Zhang M, McGranahan G and Leslie C, 2009. Validation of winter chill models using historic records of walnut phenology. Agricultural and Forest Meteorology 149, 1854-1864

Luedeling E and Brown PH, 2011. A global analysis of the comparability of winter chill models for fruit and nut trees. International Journal of Biometeorology 55, 411-421

Luedeling E, Kunz A and Blanke M, 2011. Mehr Chilling fuer Obstbaeume in waermeren Wintern? (More winter chill for fruit trees in warmer winters?). Erwerbs-Obstbau 53, 145-155

Review on chilling models in a climate change context:

Luedeling E, 2012. Climate change impacts on winter chill for temperate fruit and nut production: a review. Scientia Horticulturae 144, 218-229

The PLS method is described here:

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52.

Wold S (1995) PLS for multivariate linear modeling. In: van der Waterbeemd H (ed) Chemometric methods in molecular design: methods and principles in medicinal chemistry, vol 2. Chemie, Weinheim, pp 195-218.

Wold S, Sjostrom M, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab 58(2), 109-130.

Mevik B-H, Wehrens R, Liland KH (2011) PLS: Partial Least Squares and Principal Component Regression. R package version 2.3-0. http://CRAN.R-project.org/package0pls.

Some applications of the PLS procedure:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Yu H, Luedeling E and Xu J, 2010. Stronger winter than spring warming delays spring phenology on the Tibetan Plateau. Proceedings of the National Academy of Sciences (PNAS) 107 (51), 22151-22156.

Yu H, Xu J, Okuto E and Luedeling E, 2012. Seasonal Response of Grasslands to Climate Change on the Tibetan Plateau. PLoS ONE 7(11), e49230.

The exact procedure was used here:

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

models<-list(CP=Dynamic_Model,CU=Utah_Model,GDH=GDH)

dc<-daily_chill(stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year>2009),]),
 latitude=50.4),11,models)

Description

Converts R dates to YEARMODA format

Usage

Date2YEARMODA(Date, hours = FALSE)

Arguments

Details

Converts R date to YEARMODA

Value

YEARMODA object (e.g. 20111224 for 24th December 2011)

Author(s)

Eike Luedeling

Examples

Date2YEARMODA(YEARMODA2Date(20001205))
Date2YEARMODA(YEARMODA2Date(19901003))

Description

This function computes sunrise time, sunset time and daylength for a particular location and day of the year (Julian day). This is done using equations by Spencer (1971) and Almorox et al. (2005).

Usage

daylength(latitude, JDay, notimes.as.na = FALSE)

Arguments

Value

list with three elements Sunrise, Sunset and Daylength. For days without sunrise (polar nights), sunset and sunrise become -99 and the daylength 0. For days without sunset, sunset and sunrise are 99 and daylength 24.

Author(s)

Eike Luedeling

References

Spencer JW, 1971. Fourier series representation of the position of the Sun. Search 2(5), 172.

Almorox J, Hontoria C and Benito M, 2005. Statistical validation of daylength definitions for estimation of global solar radiation in Toledo, Spain. Energy Conversion and Management 46(9-10), 1465-1471)

Examples

daylength(latitude=50,JDay=40)
plot(daylength(latitude=35,JDay=1:365)$Daylength)

Description

Accesses the CMIP6 data of the Copernicus API via the ecmwfr package. Saves the downloaded files as .zip objects in the specified path in a subfolder with the coordinates of the downloaded area as subfolder name. You can either specify the GCMs by name or you can take all GCMs for which you downloaded climate change scenarios (model = "match_downloaded").

Usage

download_baseline_cmip6_ecmwfr(
  area,
  user,
  key,
  model = "match_downloaded",
  service = "cds",
  frequency = "monthly",
  variable = c("Tmin", "Tmax"),
  year_start = 1986,
  year_end = 2014,
  month = 1:12,
  sec_wait = 3600,
  n_try = 10,
  update_everything = FALSE,
  path_download = "cmip6_downloaded"
)

Arguments

Details

Registering for cds.climate.coperincus.eu: https://cds.climate.copernicus.eu/cdsapp#!/home

Value

NULL, the downloaded files are saved in the stated directory

Author(s)

Lars Caspersen

Examples

## Not run: 
# example with one specified GCM 
download_baseline_cmip6_ecmwfr(
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here',
    key = 'write key here',
    model = 'AWI-CM-1-1-MR',
    frequency = 'monthly', 
    variable = c('Tmin', 'Tmax'))
 
## End(Not run)

Description

Accesses the CMIP6 data of the Copernicus API via the ecmwfr package. Saves the downloaded files as .zip objects in the specified path in a subfolder with the coordinates of the downloaded area as subfolder name. You can either specify the GCMs by name, take the combinations of scenario and GCM that worked in the past (model = 'default') or you can try out all GCMs for a scenario and take the ones for which there is data (model = 'all').

Usage

download_cmip6_ecmwfr(
  scenarios,
  area,
  user,
  key,
  model = "default",
  service = "cds",
  frequency = "monthly",
  variable = c("Tmin", "Tmax"),
  year_start = 2015,
  year_end = 2100,
  month = 1:12,
  sec_wait = 3600,
  n_try = 10,
  update_everything = FALSE,
  path_download = "cmip6_downloaded"
)

Arguments

Details

Registering for cds.climate.coperincus.eu: https://cds.climate.copernicus.eu/cdsapp#!/home

Finding the user id and the key:

On the website of the Copernicus climate data store, navigate to the user profile and scroll to the bottom to "API key". There you can find the item "UID". The user id should be provided as character (within quotation marks). Just below, you can also find the key, which is also needed when using this function.

After successful registration some extra steps are needed in order to be able to download CMIP6 data. In addition to the "Terms of use of the Copernicus Climate Store" and the "Data Protection and Privacy Agreement", you also need to agree to the "CMIP6 - Data Access - Terms of Use". This needs to be done after registering. You can agree to the terms via the following link: https://cds.climate.copernicus.eu/cdsapp/#!/terms/cmip6-wps.

Alternatively, you can navigate to the terms within the Copernicus webpage. Go to "Datasets", you can find it in the upper ribbon of the main page. There you need to search for "CMIP6" using the search field and choose the first result, which is named "CMIP6 climate projections". There you need to click on "Download data" and scroll to the very bottom of the page to the field "Terms of Use". There you need to click on the button saying "Accept Terms". If you do not accept the terms the download via the API (and consequently via this function) will not be possible!

Sometimes the server is not responding in time, which can make the download fail. In such cases, after a short waiting time of 5 seconds, the request is started again. If the error reoccurs several times, the requested model will be dropped from the list of requests. By default the number of allowed repeated requests is 10. The user will get a warning if the model is dropped from the requests.

Value

NULL, the downloaded files are saved in the stated directory

Author(s)

Lars Caspersen, Antonio Picornell

Examples

## Not run: 
# example with one specified GCM 
download_cmip6_ecmwfr(
    scenarios = 'ssp126', 
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here'
    key = 'write key here',
    model = 'AWI-CM-1-1-MR',
    frequency = 'monthly', 
    variable = c('Tmin', 'Tmax'),
    year_start = 2015, 
    year_end = 2100)

# example with default combinations of scenario and GCM
download_cmip6_ecmwfr(
    scenarios = 'ssp126', 
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here'
    key = 'write key here',
    model = 'default',
    frequency = 'monthly', 
    variable = c('Tmin', 'Tmax'),
    year_start = 2015, 
    year_end = 2100)

# example with all possible combinations of scenario and GCM
# this may take a little longer
download_cmip6_ecmwfr(
    scenarios = 'ssp126', 
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here'
    key = 'write key here',
    model = 'all',
    frequency = 'monthly', 
    variable = c('Tmin', 'Tmax'),
    year_start = 2015, 
    year_end = 2100)

## End(Not run)

Description

Calculation of cumulative chill according to the Dynamic Model

This function calculates winter chill for temperate trees according to the Dynamic Model.

Chill Portions are calculated as suggested by Erez et al. (1990).

Usage

Dynamic_Model(
  HourTemp,
  summ = TRUE,
  E0 = 4153.5,
  E1 = 12888.8,
  A0 = 139500,
  A1 = 2.567e+18,
  slope = 1.6,
  Tf = 277
)

Arguments

Value

Vector of length length(HourTemp) containing the cumulative Chill Portions over the entire duration of HourTemp.

Author(s)

Eike Luedeling

References

Dynamic Model references:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

res <- Dynamic_Model(hourtemps$hourtemps$Temp)

Description

Calculation of cumulative chill according to the Dynamic Model

This function calculates winter chill for temperate trees according to the Dynamic Model.

Chill Portions are calculated as suggested by Erez et al. (1990).

Usage

DynModel_driver(
  temp,
  times,
  A0 = 139500,
  A1 = 2.567e+18,
  E0 = 4153.5,
  E1 = 12888.8,
  slope = 1.6,
  Tf = 4,
  deg_celsius = TRUE
)

Arguments

Details

This function gives idential results as Dynamic_Model for hourly temperature data, returns more details but is also a bit slower in the R code version

Value

List containint four vectors of length(temp) with elements x is the PDBF, y the accumulated chill, delta the chill portions and xs, which is $x_s=A_0/A_1\exp(-(E_0-E_1)/T)$ Portions over the entire duration of HourTemp.

Author(s)

Carsten Urbach <[email protected]>

References

Dynamic Model references:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])
hourtemps<-stack_hourly_temps(weather,latitude=50.4)
res2 <- DynModel_driver(temp=hourtemps$hourtemps$Temp)

Description

This function derives an empirical daily temperature curve from observed hourly temperature data. The mean temperature during each hour of the day is expressed as a function of the daily minimum and maximum temperature. This is done separately for each month of the year. The output is a data.frame that can then be used with the Empirical_hourly_temperatures function to generate hourly temperatures from data on daily minimum (Tmin) and maximum (Tmax) temperatures.

Usage

Empirical_daily_temperature_curve(Thourly)

Arguments

Value

data.frame containing three columns: Month (month for which coefficient applies), Hour (hour for which coefficient applies) and Prediction_coefficient (the coefficient used for empirical temperature prediction). Coefficients indicate, by what fraction of the daily temperature range the temperature during the specified hour is above the daily minimum temperature.

Author(s)

Eike Luedeling

Examples

Empirical_daily_temperature_curve(Winters_hours_gaps)

Description

This function generates hourly temperatures from daily minimum and maximum temperatures, based on an empirical relationship of these two daily temperature extremes with the hourly temperature. Usually, this relationship will have been determined with the
Empirical_daily_temperature_curve function.

Usage

Empirical_hourly_temperatures(Tdaily, empi_coeffs)

Arguments

Value

data.frame containing all columns of the Tdaily dataset, but also the columns Hour and Temp, for the hour of the day and the predicted temperature, respectively.

Author(s)

Eike Luedeling

Examples

coeffs<-Empirical_daily_temperature_curve(Winters_hours_gaps)
Winters_daily<-make_all_day_table(Winters_hours_gaps, input_timestep="hour")
Empirical_hourly_temperatures(Winters_daily,coeffs)

Description

Opens the downloaded .zip files and returns the CMIP6 climate projections for specified locations .

Usage

extract_cmip6_data(
  stations,
  variable = c("Tmin", "Tmax"),
  download_path = "cmip6_downloaded",
  keep_downloaded = TRUE
)

Arguments

Value

named list of data.frames. Element names follow the syntax 'SSP'_'GCM', where SSP is the shared socioeconomic pathway and GCM is the global climate model that generated the weather data. The data.frames contain the extracted values for the requested locations.

Author(s)

Lars Caspersen

Examples

## Not run: 
scenario<-c("ssp126", "ssp245", "ssp370", "ssp585")

download_cmip6_ecmwfr(scenario,
                         key = 'your-key-here'
                         user = 'your-user-name-here',
                         area =  c(52, -7, 33, 8) )


station <- data.frame(
    station_name = c('Zaragoza', 'Klein-Altendorf', 'Sfax', 'Cieza',
        'Meknes', 'Santomera'),
    longitude = c(-0.88,  6.99, 10.75, -1.41, -5.54, -1.05),
    latitude = c(41.65, 50.61, 34.75, 38.24, 33.88, 38.06))

extracted <- extract_cmip6_data(
    stations = station)
    
scenario <- gen_rel_change_scenario(
    extracted, years_local_weather = c(1992, 2021))


## End(Not run)

Description

For a vector of character strings, identify elements between shared leading and/or trailing substrings, e.g. for a vector such as c("XXX01YYY",XXX02YYY") extract the numbers.

Usage

extract_differences_between_characters(strings)

Arguments

Value

vector of strings similar to the input vector but without shared leading and trailing characters.

Author(s)

Eike Luedeling

Examples

extract_differences_between_characters(c("Temp_01","Temp_02","Temp_03"))
  extract_differences_between_characters(c("Temp_01_Tmin","Temp_02_Tmin","Temp_03_Tmin"))
  extract_differences_between_characters(c("a","b"))

Description

Temperature data is often available in gridded format, and records for particular points must be extracted for work on site-specific issues (such as chill calculation). This function implements this, for certain types of gridded data.

Usage

extract_temperatures_from_grids(
  coordinates,
  grid_format,
  grid_specifications,
  scenario_year = NA,
  reference_year = NA,
  scenario_type = NA,
  labels = NA,
  temperature_check_args = NULL
)

Arguments

Details

The following climate data formats are supported: "AFRICLIM" - data downloaded from https://www.york.ac.uk/environment/research/kite/resources/; "CCAFS" - data downloaded from http://ccafs-climate.org/data_spatial_downscaling/; "WorldClim" - data downloaded from http://www.worldclim.org/. All these databases provide separate zipped files for monthly minimum and monthly maximum temperatures, but they differ slightly in format and structure. If you want to see additional formats included, please send me a message.

Value

temperature scenario object extracted from the grids, consisting of the following elements: 'data' = a data frame with n_intervals elements containing the absolute or relative temperature information. 'reference_year' = the year the scenario is representative of. 'scenario_type' = the scenario type ('absolute' or 'relative'); 'labels' = and elements attached to the input temperature_scenario as an element names 'labels'.

The function generates errors, when problems arise.

Author(s)

Eike Luedeling

Examples

coordinates<-c(10.6082,34.9411)
 # grid_specifications<-list(base_folder="D:/DATA/AFRICLIM/GeoTIFF_30s/future_scenarios/",
 #                           minfile="tasmin_rcp45_2055_CCCma-CanESM2_CCCma-CanRCM4_wc30s.zip",
 #                           maxfile="tasmax_rcp45_2055_CCCma-CanESM2_CCCma-CanRCM4_wc30s.zip")
                            
 # extract_temperatures_from_grids(coordinates,grid_format="AFRICLIM",grid_specifications,
 #    scenario_type="relative",scenario_year=2055)
                 
 # grid_specifications<-list(base_folder="D:/DATA/CCAFS_climate/",
 #                           minfile="bcc_csm1_1_rcp2_6_2030s_tmin_30s_r1i1p1_b4_asc.zip",
 #                           maxfile="bcc_csm1_1_rcp2_6_2030s_tmax_30s_r1i1p1_b4_asc.zip")
 #temps<-extract_temperatures_from_grids(coordinates,grid_format="CCAFS",grid_specifications,
 #                                       scenario_type="relative",scenario_year=2035)

Description

This function attempts to remove erroneous temperature readings. This is tricky because of the wide range of errors that can occur, so this isn't necessarily sufficient for problems of particular records.

Usage

filter_temperatures(
  temp_file,
  remove_value = NA,
  running_mean_filter = NA,
  running_mean_length = 3,
  min_extreme = NA,
  max_extreme = NA,
  max_missing_in_window = 1,
  missing_window_size = 9
)

Arguments

Value

filtered temperature dataset, from which records identified as erroneous were removed.

Author(s)

Eike Luedeling

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2009),])

hourtemps<-stack_hourly_temps(weather, latitude=50.4)

filtered<-filter_temperatures(hourtemps$hourtemps,remove_value=-99,
  running_mean_filter=3)

Description

This function identifies and interpolates gaps in daily weather records

Usage

fix_weather(
  weather,
  start_year = 0,
  end_year = 3000,
  start_date = 1,
  end_date = 366,
  columns = c("Tmin", "Tmax"),
  end_at_present = TRUE
)

Arguments

Details

This function produces a complete record containing all dates between the 1st day of the start year and the last day of the end year (unless the first/last day of the record is after/before these dates - in that case the record is not extended). The values for the columns specified by the columns attribute are linearly interpolated. Missing values during the period indicated by start_date and end_date are added up and summarized in a quality control table.

Value

list with two elements: weather: contains the interpolated weather record QC: contains the quality control data.frame, which summarizes missing days, incomplete days (days on which any value is missing), and percentage completeness.

Author(s)

Eike Luedeling

Examples

fix_weather(KA_weather,2000,2010)

#use a subset of the KA_weather dataset and add an additional day after a gap
KA_weather_gap<-rbind(KA_weather,c(Year=2011,Month=3,Day=3,Tmax=26,Tmin=14)) 
#fill in the gaps
fix_weather(KA_weather_gap, 1990,2011,300,100)

#fix_weather(KA_weather)

Description

This function calculates heat for temperate trees according to the Growing Degree Day Model. Note that the calculuation differs slightly from the original, in which it is based on daily temperature extremes only. This equation here works with hourly temperatures. The normal GDD equation is GDD=(Tmax-Tmin)/2-Tbase, with Tmax=30 for Tmax>30, and Tmin=10 for Tmin<10. Tbase is a species-specific base temperature. The first part of the equation is the arithmetic mean of daily temperature extremes. In the present equation, this is replaced by Thourly/24 for each hourly temperature value. If chillR was using a triangular daily temperature curve, the result would be the same for both equations. Since chillR uses a sine function for daytime warming and a logarithmic decay function for nighttime cooling, however, there will be a slight deviation. This could be handled by defining a function the runs with daily weather data. chillR doesn't currently have this capability, since its primary focus is on metrics that require hourly data.

Usage

GDD(HourTemp, summ = TRUE, Tbase = 5)

Arguments

Details

Growing Degree Hours are calculated as suggested by Anderson et al. (1986).

Value

Vector of length length(HourTemp) containing the cumulative Growing Degree Days over the entire duration of HourTemp.

Author(s)

Eike Luedeling

References

Growing Degree Days reference:

http://agron-www.agron.iastate.edu/Courses/agron212/Calculations/GDD.htm

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

GDD(hourtemps$hourtemps$Temp)

Description

This function calculates heat for temperate trees according to the Growing Degree Hours Model.

Usage

GDH(HourTemp, summ = TRUE)

Arguments

Details

Growing Degree Hours are calculated as suggested by Anderson et al. (1986).

Value

Vector of length length(HourTemp) containing the cumulative Growing Degree Hours over the entire duration of HourTemp.

Author(s)

Eike Luedeling

References

Growing Degree Hours reference:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

GDH(hourtemps$hourtemps$Temp)

Description

This function calculates heat for temperate trees according to the Growing Degree Hours Model.

Usage

GDH_model(HourTemp, summ = TRUE)

Arguments

Details

Growing Degree Hours are calculated as suggested by Anderson et al. (1986).

Value

Vector of length length(HourTemp) containing the cumulative Growing Degree Hours over the entire duration of HourTemp.

Author(s)

Eike Luedeling

References

Growing Degree Hours reference:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2006),])

hourtemps<-stack_hourly_temps(weather,latitude=50.4)

GDH_model(hourtemps$hourtemps$Temp)

Description

Takes the extracted CMIP6 data and returns climate change scenarios, which can then be used to generate weather data.

Usage

gen_rel_change_scenario(
  downloaded_list,
  scenarios = c(2050, 2085),
  reference_period = c(1986:2014),
  future_window_width = 30
)

Arguments

Value

data.frame for the calculated relative change scenarios, all locations, SSPs, timepoints, GCMs combined

Author(s)

Lars Caspersen

Examples

## Not run: 
download_cmip6_ecmwfr(scenario = 'ssp1_2_6', 
                      area = c(55, 5.5, 47, 15.1),
                      user = 'write user id here',
                      key = 'write key here',
                      model = 'AWI-CM-1-1-MR',
                      frequency = 'monthly', 
                      variable = c('Tmin', 'Tmax'),
                      year_start = 2015, 
                      year_end = 2100)
                      
download_baseline_cmip6_ecmwfr(
    area = c(55, 5.5, 47, 15.1),
    user = 'write user id here',
    key = 'write key here',
    model = 'AWI-CM-1-1-MR',
    frequency = 'monthly', 
   
station <- data.frame(
      station_name = c('Zaragoza', 'Klein-Altendorf', 'Sfax',
      'Cieza', 'Meknes', 'Santomera'),
      longitude = c(-0.88,  6.99, 10.75, -1.41, -5.54, -1.05),
      latitude = c(41.65, 50.61, 34.75, 38.24, 33.88, 38.06))
      
extracted <- extract_cmip6_data(stations = station)

gen_rel_change_scenario(extracted)


## End(Not run)

Description

Identify the hours, days or months in a (monthly, daily or hourly) temperature dataset that belong to a particular season. Seasons are defined according to the 'mrange' argument, which specifies the start and end month of the season. The 'years' argument specifies the year, in which the dormancy season of interest ends.

Usage

genSeason(temps, mrange = c(8, 6), years)

Arguments

Description

Generates a list with data.frame elements for each season.

Usage

genSeasonList(temps, mrange = c(8, 6), years)

Arguments

Value

Returns a list of data frames. Each element of the list corresponds to one season. The 'data.frame' for each year has named columns 'Temp', 'JDay' and 'Year'.

Description

When looking at multi-year phenology records, it is normally obvious in which year bloom occurred last. Determining this with an automated procedure, however, is a bit tricky, when the range of phenological dates spans across a calendar year transition. This function finds the latest phenological date of the record. This is the date before the longest phenological date gap.

Usage

get_last_date(dates, first = FALSE)

Arguments

Value

the latest (earliest) date of the series, under the assumption that the longest period without bloom can be interpreted as separating the phenological seasons. This should be a reasonable assumption in most cases.

Author(s)

Eike Luedeling

Examples

get_last_date(c(1,3,6,8,10,25))
get_last_date(c(345,356,360,365,2,5,7,10))
get_last_date(c(345,356,360,365,2,5,7,10),first=TRUE)

Description

This function retrieves either a list of nearby weather stations for a specified point location, or it downloads weather data for a specific weather station.

Usage

get_weather(
  location,
  time_interval = NA,
  database = "UCIPM",
  station_list = NULL,
  stations_to_choose_from = 25,
  end_at_present = TRUE
)

Arguments

Details

weather databases, from which chillR can download data: NOAA NCDC Global Summary of the Day - "GSOD" (https://data.noaa.gov/dataset/global-surface-summary-of-the-day-gsod)

California Irrigation Management Information System (CIMIS) - "CIMIS" (http://www.cimis.water.ca.gov/)

University of California Integrated Pest Management (UCIPM) - "UCIPM" (http://ipm.ucdavis.edu/WEATHER/)

several formats are possible for specifying the location vector, which can consist of either two or three coordinates (it can include elevation). Possible formats include c(1,2,3), c(1,2), c(x=1,y=2,z=3), c(lat=2,long=1,elev=3). If elements of the vector are not names, they are interpreted as c(Longitude, Latitude, Elevation).

The 'chillRCode' is generated by this function, when it is run with geographic coordinates as location inputs. In the list of nearby stations that is returned then, the chillRCode is provided and can then be used as input for running the function in 'downloading' mode. For downloading the data, use the same call as before but replace the location argument with the chillRCode.

Value

The output depends on how the location is provided. If it is a coordinate vector, the function returns a list of station_to_choose_from weather stations that are close to the specified location. This list also contains information about how far away these stations are (in km), how much the elevation difference is (if elevation is specified; in m) and how much overlap there is between the data contained in the database and the time period specified by time_interval.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

see the documentation of the handler functions (e.g. handle_ucipm) for details.

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

#stat_list<-handle_gsod(action="list_stations",location=c(x=-122,y=38.5),
#  time_interval=c(2002,2002))
#the line above takes longer to run than CRAN allows for examples. The line below therefore
#generates an abbreviated stat_list that allows running the code.
stat_list<-data.frame(chillR_code=c("724828_99999","724828_93241","720576_174"),
   Lat=c(38.383,38.378,38.533),Long=c(-121.967,-121.958,-121.783),
   BEGIN=c(20010811,20060101,20130101),END=c(20051231,20160110,20160109))

#gw<-get_weather(location="724828_93241",time_interval=c(2012,2012),database="GSOD",
#   station_list = stat_list)

#stat_list<-get_weather(location=c(lat=50,lon=10,ele=150),time_interval=c(2001,2001),
#  database="UCIPM")
#chillRcode<-stat_list[which(stat_list$Perc_interval_covered==
#max(stat_list$Perc_interval_covered)),"chillR_code"][1]
  #after the first few lines here, the code should be "CEDARVIL.C"
#gw<-get_weather(location="CEDARVIL.C",time_interval=c(2001,2001),database="UCIPM")
#weather<-weather2chillR(gw,"GSOD")
#make_chill_plot(tempResponse(stack_hourly_temps(fix_weather(weather))),
#                "Chill_Portions",start_year=2005,end_year=2011,metriclabel="Chill Portions")

Description

This function is a wrapper for the getClimateWizardData function to access climate scenario data for a location of interest. Climate model runs are queried and data returned and summarized according to the specified parameters. A number of metrics are available for several climate models, which are listed in https://github.com/CIAT-DAPA/climate_wizard_api. This function can download data for multiple climate scenarios, saving users the effort to retrieve them separately.

Usage

getClimateWizard_scenarios(
  coordinates,
  scenarios,
  start_years,
  end_years,
  baseline = c(1950, 2005),
  metric = "monthly_min_max_temps",
  GCMs = "all"
)

Arguments

Details

Note that this function lacks quality checks. If something goes wrong, you may consider checking individual scenarios with the getClimateWizardData function.

Value

data.frame containing the requested information.

Author(s)

Eike Luedeling

References

Girvetz E, Ramirez-Villegas J, Navarro C, Rodriguez C, Tarapues J, undated. ClimateWizard REST API for querying climate change data. https://github.com/CIAT-DAPA/climate_wizard_api

Examples

#example is #d out, because of runtime issues.
#getC<-getClimateWizard_scenarios(coordinates=c(longitude=6.99,latitude=50.62),
#                                scenarios=c("rcp85","rcp45"),
#                                start_years=c(2070,2035),
#                                end_years=c(2100,2065),
#                                metric=c("monthly_tmean"),
#                                GCMs=c("all"))

Description

This function makes use of an API provided by the International Center for Tropical Agriculture (CIAT) to access climate scenario data for a location of interest. Climate model runs are queried and data returned and summarized according to the specified parameters. A number of metrics are available for several climate models, which are listed in the API repository. Refer to this document for details on what can be downloaded. This function provides the additional option of automatically retrieving all data referring to changes in daily temperature extremes (by month), by setting the “'metric“' parameter to "monthly_min_max_temps". It also offers the option to automatically obtain data for all climate models included in the database (as of January 2018).

Usage

getClimateWizardData(
  coordinates,
  scenario,
  start_year,
  end_year,
  baseline = c(1950, 2005),
  metric = "monthly_min_max_temps",
  GCMs = "all",
  temperature_generation_scenarios = FALSE
)

Arguments

Value

data.frame containing the requested information.

Author(s)

Eike Luedeling

References

Girvetz E, Ramirez-Villegas J, Navarro C, Rodriguez C, Tarapues J, undated. ClimateWizard REST API for querying climate change data. https://github.com/CIAT-DAPA/climate_wizard_api

Examples

# the example is #d out, since the download request sometimes times out, and that
# causes problems with CRAN approval of the package

# getClimateWizardData(coordinates=c(longitude=10.613975,latitude=34.933439),
#   scenario="rcp45", start_year=2020, end_year=2050,
#   metric=c("CD18","R02"), GCMs=c("bcc-csm1-1","BNU-ESM"))

Description

This function can do three things related to the California Irrigation Management Information System ("CIMIS") database: 1. it can list stations that are close to a specified position (geographic coordinates) 2. it can retrieve weather data for a named weather station 3. it can 'clean' downloaded data, so that they can easily be used in chillR Which of these functions is carried out depends on the action argument.

Usage

handle_cimis(
  action,
  location = NA,
  time_interval = NA,
  station_list = NULL,
  stations_to_choose_from = 25,
  drop_most = TRUE,
  end_at_present = TRUE
)

Arguments

Details

This function can run independently, but it is also called by the get_weather and weather2chillR functions, which some users might find a bit easier to handle.

The CIMIS dataset is described here: http://www.cimis.water.ca.gov/

Under the 'list_stations' mode, several formats are possible for specifying the location vector, which can consist of either two or three coordinates (it can include elevation). Possible formats include c(1,2,3), c(1,2), c(x=1,y=2,z=3), c(lat=2,long=1,elev=3). If elements of the vector are not names, they are interpreted as c(Longitude, Latitude, Elevation).

The 'chillRCode' is generated by this function, when it is run with geographic coordinates as location inputs. In the list of nearby stations that is returned then, the chillRCode is provided and can then be used as input for running the function in 'downloading' mode. For downloading the data, use the same call as before but replace the location argument with the chillRCode.

Value

The output depends on the action argument. If it is 'list_stations', the function returns a list of station_to_choose_from weather stations that are close to the specified location. This list also contains information about how far away these stations are (in km), how much the elevation difference is (if elevation is specified; in m) and how much overlap there is between the data contained in the database and the time period specified by time_interval. If action is 'download_weather' the output is a list of two elements: 1. database="CIMIS" 2. the downloaded weather record, extended to the full duration of the specified time interval. If action is a weather data.frame or a weather record downloaded with this function (in 'download_weather' mode), the output is the same data in a format that is easy to use in chillR. If drop_most was set to TRUE, most columns are dropped.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

Past CIMIS data is provided to the public as compressed data files of annual data, which contain data for all stations for the respective years. The same strategy was followed for monthly data of the past year. This means that in order to get to the records for one given station, it is necessary to download data for all stations first, before extracting weather for the station of interest. This means that downloads take a lot longer than one might expect, and the downloaded data volume is a multiple of what is really of interest.

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

# the example is #d out, since the download request sometimes times out, and that
# causes problems with CRAN approval of the package

# handle_cimis(action = "list_stations",
#              location = c(x = -122, y = 38.5),
#              time_interval = c(2012, 2012))

# stat_list <- data.frame("Station Number" = c("119", "139", "6"),
#                         Latitude = c(38.49500, 38.50126, 38.53569),
#                         Longitude = c(-122.0040, -121.9785, -121.7764),
#                         Start_date =c("1993-08-21 UTC", "1998-06-15 UTC", "1982-07-17 UTC"),
#                         End_date = c("1995-01-25", "2016-03-06", "2016-03-06"))

# gw <- handle_cimis(action = "download_weather",
#                    location = "6",
#                    time_interval = c(1982, 1982), 
#                    station_list = stat_list)
 
# weather <- handle_cimis(gw)

# make_chill_plot(tempResponse(stack_hourly_temps(fix_weather(weather)),
#                 Start_JDay = 300, End_JDay = 50), 
#                 "Chill_Portions", start_year = 2010, end_year = 2012,
#                 metriclabel = "Chill Portions", misstolerance = 50)

Description

This function accesses the Deutscher Wetterdienst database and allows to:

• 1) list a number of weather stations that are close to a specific position (geographic coordinates)

• 2) obtain weather data for one or more weather stations through the station ID

• 3) 'clean' and 'format' downloaded data, so the records can easily be used in other chillR functions

Usage

handle_dwd(
  action,
  location = NA,
  time_interval = c(19160101, Date2YEARMODA(Sys.Date())),
  station_list = NULL,
  stations_to_choose_from = 25,
  drop_most = TRUE,
  end_at_present = TRUE,
  add.DATE = FALSE,
  quiet = FALSE,
  add_station_name = FALSE
)

Arguments

Value

If action = 'list_stations', the function returns a data frame with 'stations_to_choose_from' rows and 9 columns. This data frame contains information about the weather stations (Latitude, Longitude, among others). If action = 'download_weather', the function returns a list of length according to the length of the location parameter. Each list elements is a data frame containing the data downloaded from the database. If the action is provided with the list generated by the function in the downloading mode, the function will return a list of data frames structured according to the chillR format. If drop_most is set to TRUE, the function will keep only the most relevant variables for standard chillR analyses.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

Author(s)

Eduardo Fernandez and Eike Luedeling

References

Fernandez, E., Whitney, C., and Luedeling, E. 2020. The importance of chill model selection - A multi-site analysis. European Journal Of Agronomy 119: 126103

Examples

# The following lines may take longer than required to pass the
# CRAN checks. Please, un-comment them to run the example

# stations <- handle_dwd(action = "list_stations", 
#                        location = c(latitude = 53.5373, longitude = 9.6397),
#                        time_interval = c(20000101, 20101231),
#                        stations_to_choose_from = 25)
                     
# data <- handle_dwd(action = "download_weather",
#                     location = stations[1 : 3, "Station_ID"],
#                     time_interval = c(20000101, 20020601),
#                     stations_to_choose_from = 25,
#                     station_list = stations,
#                     drop_most = TRUE, 
#                     add.DATE = FALSE,
#                     quiet = TRUE,
#                     add_station_name = FALSE)

# data_modified <- handle_dwd(data, add.DATE = TRUE, drop_most = TRUE)

Description

This function is deprecated and will disappear soon. Please refer to the handle_dwd function for the most current functionality.

Usage

handle_dwd_old(
  action,
  location = NA,
  time_interval = c(19160101, Date2YEARMODA(Sys.Date())),
  station_list = NULL,
  stations_to_choose_from = 25,
  drop_most = TRUE,
  end_at_present = TRUE,
  add.DATE = FALSE,
  quiet = FALSE,
  add_station_name = FALSE
)

Arguments

Details

This function accesses the Deutscher Wetterdienst database and allows to:

• 1) list a number of weather stations that are close to a specific position (geographic coordinates)

• 2) obtain weather data for one or more weather stations through the station ID

• 3) 'clean' and 'format' downloaded data, so the records can easily be used in other chillR functions

Value

If action = 'list_stations', the function returns a data frame with 'stations_to_choose_from' rows and 9 columns. This data frame contains information about the weather stations (Latitude, Longitude, among others). If action = 'download_weather', the function returns a list of length according to the length of the location parameter. Each list, is a list of two elements; a data frame containing the data downloaded from the database and character string representing the respective database ('dwd'). If the action is provided with the list generated by the function in the downloading mode, the function will return a list of data frames structured according to the chillR format. If drop_most is set to TRUE, the function will keep only the relevant variables.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

Author(s)

Eduardo Fernandez and Eike Luedeling

References

Fernandez, E., Whitney, C., and Luedeling, E. 2020. The importance of chill model selection - A multi-site analysis. European Journal Of Agronomy 119: 126103

Examples

# The following lines may take longer than required to pass the
# CRAN checks. Please, un-comment them to run the example

# stations <- handle_dwd_old(action = "list_stations", 
#                        location = c(latitude = 53.5373, longitude = 9.6397),
#                        time_interval = c(20000101, 20101231),
#                        stations_to_choose_from = 25)
                     
# data <- handle_dwd_old(action = "download_weather",
#                     location = stations[1 : 3, "Station_ID"],
#                     time_interval = c(20000101, 20020601),
#                     stations_to_choose_from = 25,
#                     station_list = stations,
#                     drop_most = TRUE, 
#                     add.DATE = FALSE,
#                     quiet = TRUE,
#                     add_station_name = FALSE)

# data_modified <- handle_dwd_old(data, add.DATE = TRUE, drop_most = TRUE)

Description

This function can do four things related to the Global Summary of the Day ("GSOD") database from the National Climatic Data Centre (NCDC) of the National Oceanic and Atmospheric Administration (NOAA):

• 1. It can list stations that are close to a specified position (geographic coordinates).

• 2. It can retrieve weather data for a named weather station (or a vector of multiple stations). For the name, the chillRcode from the list returned by the list_stations operation should be used.

• 3. It can 'clean' downloaded data (for one or multiple stations), so that they can easily be used in chillR

• 4. It can delete the downloaded intermediate weather files from the machine

  Which of these functions is carried out depends on the action argument.

This function can run independently, but it is also called by the get_weather and weather2chillR functions, which some users might find a bit easier to handle.

Usage

handle_gsod(
  action,
  location = NULL,
  time_interval = c(1950, 2020),
  stations_to_choose_from = 25,
  end_at_present = FALSE,
  add.DATE = FALSE,
  update_station_list = FALSE,
  path = "climate_data",
  update_all = FALSE,
  clean_up = NULL,
  override_confirm_delete = FALSE,
  max_distance = 150,
  min_overlap = 0,
  verbose = "normal"
)

Arguments

Details

The GSOD database is described here: https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00516

under the 'list_stations' mode, several formats are possible for specifying the location vector, which can consist of either two or three coordinates (it can include elevation). Possible formats include c(1, 2, 3), c(1, 2), c(x = 1, y = 2, z = 3), c(lat = 2, long = 1, elev = 3). If elements of the vector are not names, they are interpreted as c(Longitude, Latitude, Elevation).

The 'chillRCode' is generated by this function, when it is run with geographic coordinates as location inputs. In the list of nearby stations that is returned then, the chillRCode is provided and can then be used as input for running the function in 'downloading' mode. For downloading the data, use the same call as before but replace the location argument with the chillRCode.

Value

The output depends on the action argument. If it is 'list_stations', the function returns a list of station_to_choose_from weather stations that are close to the specified location. This list also contains information about how far away these stations are (in km), how much the elevation difference is (if elevation is specified; in m) and how much overlap there is between the data contained in the database and the time period specified by time_interval. If action is 'download_weather' the output is a list of the downloaded weather record, extended to the full duration of the specified time interval. If the location input was a vector of stations, the output will be a list of such objects. If action is a weather data.frame or a weather record downloaded with this function (in 'download_weather' mode), the data structure remains in the same, but the data are processed for easy use with chillR. If drop_most was set to TRUE, most columns are dropped. If the location input was a list of weather datasets, all elements of the list will be processed. **IMPORTANT NOTE:** as of chillR version 0.73, the output format no longer contains a list element that specifies the database name, because this has been considered confusing (and annoying) by various users. This means, however, that some earlier calls to results from the handle_gsod function may produce errors now. Also note that a few parameters, station_list, drop_most, quiet, add_station_name are no longer needed due to some reworking of the function's mechanisms. After careful consideration, we decided to drop these parameters entirely, which may lead to some downward compatibility problems. Apologies for any inconvenience caused by this transition. If you want to keep using the previous function (which is much slower), feel free to adopt the deprecated handle_gsod_old function - but note that this will no longer be updated and may disappear eventually.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

For many places, the GSOD database is quite patchy, and the length of the record indicated in the summary file isn't always very useful (e.g. there could only be two records for the first and last date). Files are downloaded by year, so if we specify a long interval, this may take a bit of time.

Author(s)

Adrian Fülle, Lars Caspersen, Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

#coordinates of Bonn
long <- 7.0871843
lat <- 50.7341602

#get a list of close-by weather stations
# stationlist <-
#   handle_gsod(action = "list_stations",
#               time_interval = c(1995,2000),
#               location = c(long,lat))

#download data
# test_data <-
#   handle_gsod(action = "download_weather",
#               time_interval = c(1995,2000),
#               location = stationlist$chillR_code[c(1,2)])
# 
# format downloaded data
# test_data_clean <- handle_gsod(action = test_data)

## data deletion on disk for clean_up

# functions will ask for confirmation in the console - 'y' for yes to
# confirm deletion, anything else cancels the deletion

# handle_gsod(action = "delete",
#             clean_up = "all",
#             override_confirm_delete = TRUE)

Description

This function is deprecated, but it will be retained for a few generations of updates. Its functionality has been fully replaced by the new version of the handle_gsod function, which does the same job, but much faster. That's probably the function you really want to use.

This function can do three things related to the Global Summary of the Day ("GSOD") database from the National Climatic Data Centre (NCDC) of the National Oceanic and Atmospheric Administration (NOAA):

• 1. It can list stations that are close to a specified position (geographic coordinates).

• 2. It can retrieve weather data for a named weather station (or a vector of multiple stations). For the name, the chillRcode from the list returned by the list_stations operation should be used.

• 3. It can 'clean' downloaded data (for one or multiple stations), so that they can easily be used in chillR

  Which of these functions is carried out depends on the action argument.

This function can run independently, but it is also called by the get_weather and weather2chillR functions, which some users might find a bit easier to handle.

Usage

handle_gsod_old(
  action,
  location = NA,
  time_interval = NA,
  station_list = NULL,
  stations_to_choose_from = 25,
  drop_most = TRUE,
  end_at_present = TRUE,
  add.DATE = TRUE,
  quiet = FALSE,
  add_station_name = FALSE
)

Arguments

Details

The GSOD database is described here: https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00516

under the 'list_stations' mode, several formats are possible for specifying the location vector, which can consist of either two or three coordinates (it can include elevation). Possible formats include c(1, 2, 3), c(1, 2), c(x = 1, y = 2, z = 3), c(lat = 2, long = 1, elev = 3). If elements of the vector are not names, they are interpreted as c(Longitude, Latitude, Elevation).

The 'chillRCode' is generated by this function, when it is run with geographic coordinates as location inputs. In the list of nearby stations that is returned then, the chillRCode is provided and can then be used as input for running the function in 'downloading' mode. For downloading the data, use the same call as before but replace the location argument with the chillRCode.

Value

The output depends on the action argument. If it is 'list_stations', the function returns a list of station_to_choose_from weather stations that are close to the specified location. This list also contains information about how far away these stations are (in km), how much the elevation difference is (if elevation is specified; in m) and how much overlap there is between the data contained in the database and the time period specified by time_interval. If action is 'download_weather' the output is a list of two elements: 1. database="GSOD" 2. the downloaded weather record, extended to the full duration of the specified time interval. If the location input was a vector of stations, the output will be a list of such objects. If action is a weather data.frame or a weather record downloaded with this function (in 'download_weather' mode), the output is the same data in a format that is easy to use in chillR. If drop_most was set to TRUE, most columns are dropped. If the location input was a list of weather datasets, all elements of the list will be processed.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

For many places, the GSOD database is quite patchy, and the length of the record indicated in the summary file isn't always very useful (e.g. there could only be two records for the first and last date). Files are downloaded by year, so if we specify a long interval, this may take a bit of time.

Author(s)

Eike Luedeling and Eduardo Fernandez

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

# List the near weather stations
# stat_list <- handle_gsod_old(action = "list_stations",
#                          location = c(x = -122, y = 38.5),
#                          time_interval = c(2002, 2002))

# the line above takes longer to run than CRAN allows for examples.
# The line below therefore
# generates an abbreviated stat_list that allows running the code.

# stat_list <- data.frame(chillR_code = c("724828_99999",
#                                         "724828_93241",
#                                         "720576_174"),
#                         STATION.NAME = c("NUT TREE",
#                                          "NUT TREE AIRPORT",
#                                          "UNIVERSITY AIRPORT"),
#                         Lat = c(38.383, 38.378, 38.533),
#                         Long = c(-121.967, -121.958, -121.783),
#                         BEGIN = c(20010811, 20060101, 20130101),
#                         END = c(20051231, 20160110, 20160109))

# gw <- handle_gsod_old(action = "download_weather",
#                   location = "724828_93241",
#                   time_interval = c(2010, 2012),
#                   station_list = stat_list,
#                   quiet = TRUE)

# weather <- handle_gsod_old(gw, add.DATE = FALSE)[[1]]$weather

# make_chill_plot(tempResponse(stack_hourly_temps(fix_weather(weather)),
#                              Start_JDay = 300, End_JDay = 50),
#                 "Chill_Portions", start_year = 2010,
#                 end_year = 2012, metriclabel = "Chill Portions",
#                 misstolerance = 50)

Description

This function can do three things related to the University of California Integrated Pest Management (UCIPM) database: 1. it can list stations that are close to a specified position (geographic coordinates) 2. it can retrieve weather data for a named weather station 3. it can 'clean' downloaded data, so that they can easily be used in chillR Which of these functions is carried out depends on the action argument.

Usage

handle_ucipm(
  action,
  location = NA,
  time_interval = NA,
  station_list = california_stations,
  stations_to_choose_from = 25,
  drop_most = TRUE,
  end_at_present = TRUE
)

Arguments

Details

This function can run independently, but it is also called by the get_weather and weather2chillR functions, which some users might find a bit easier to handle.

the UCIPM dataset is described here: http://ipm.ucdavis.edu/WEATHER/

under the 'list_stations' mode, several formats are possible for specifying the location vector, which can consist of either two or three coordinates (it can include elevation). Possible formats include c(1,2,3), c(1,2), c(x=1,y=2,z=3), c(lat=2,long=1,elev=3). If elements of the vector are not names, they are interpreted as c(Longitude, Latitude, Elevation).

The 'chillRCode' is generated by this function, when it is run with geographic coordinates as location inputs. In the list of nearby stations that is returned then, the chillRCode is provided and can then be used as input for running the function in 'downloading' mode. For downloading the data, use the same call as before but replace the location argument with the chillRCode.

Value

The output depends on the action argument. If it is 'list_stations', the function returns a list of station_to_choose_from weather stations that are close to the specified location. This list also contains information about how far away these stations are (in km), how much the elevation difference is (if elevation is specified; in m) and how much overlap there is between the data contained in the database and the time period specified by time_interval. If action is 'download_weather' the output is a list of two elements: 1. database="CIMIS" 2. the downloaded weather record, extended to the full duration of the specified time interval. If action is a weather data.frame or a weather record downloaded with this function (in 'download_weather' mode), the output is the same data in a format that is easy to use in chillR. If drop_most was set to TRUE, most columns are dropped.

Note

Many databases have data quality flags, which may sometimes indicate that data aren't reliable. These are not considered by this function!

The station list provided by the UC IPM database doesn't contain geographic positions of the stations, which can only be accessed by station-specific websites. This function will access this information only if it was not given on the website in early 2016. Station information based on a download at that time is stored in the california_station dataset included in chillR. This was done to reduce the run time for the handle_ucipm function. It will probably be okay for the foreseeable future (stations don't change very quickly). A new version of this table can be produces with the make_california_UCIPM_station_list() function.

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

# All examples are disabled, because the database is sometimes unavailable. This then generates
# an error when R runs its package functionality checks. To run the examples, remove the # mark,
# before running the code.
#
#handle_ucipm(action="list_stations",location=c(x=-122,y=38.5),time_interval=c(2012,2012))
#gw<-handle_ucipm(action="download_weather",location="WINTERS.A",time_interval=c(2012,2012))
#weather<-handle_ucipm(gw)$weather
#make_chill_plot(tempResponse(stack_hourly_temps(fix_weather(weather)),Start_JDay=300,End_JDay=50),
#                "Chill_Portions",start_year=2010,end_year=2012,metriclabel="Chill Portions",
#                misstolerance = 50)

Description

Compares all elements of a vector of numbers or character strings and returns TRUE if they are all the same, FALSE otherwise.

Usage

identify_common_string(strings, leading = TRUE)

Arguments

Value

if there is a leading (if leading==TRUE) or trailing (if leading==FALSE) string that all elements of strings have in common, this string is returned; NA otherwise.

Author(s)

Eike Luedeling

Examples

identify_common_string(c("Temp_01","Temp_02","Temp_03"))
  identify_common_string(c("Temp_01","Temp_02","Temp_03"),leading=FALSE)
  identify_common_string(c("file1.csv","file2.csv","file3.csv"),leading=FALSE)

Description

This function linearly interpolates gaps in data series, such as daily temperature records.

Usage

interpolate_gaps(x)

Arguments

Details

The function returns a list with two elements: interp is a new vector, in which all gaps in x have been linearly interpolated. missing is a second vector, which contains information on which values were filled in by interpolation.

Value

Author(s)

Eike Luedeling

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

weather<-make_all_day_table(KA_weather)
Tmin_int<-interpolate_gaps(KA_weather[,"Tmin"])
weather[,"Tmin"]<-Tmin_int$interp
weather[,"Tmin_interpolated"]<-Tmin_int$missing

Tmax_int<-interpolate_gaps(KA_weather[,"Tmax"])
weather[,"Tmax"]<-Tmax_int$interp
weather[,"Tmax_interpolated"]<-Tmax_int$missing

#this function is integrated into the fix_weather function, but it can also be run on its own.

Description

Using idealized temperature curves for guidance, this function interpolated hourly temperature data.

Usage

interpolate_gaps_hourly(
  hourtemps,
  latitude = 50,
  daily_temps = NULL,
  interpolate_remaining = TRUE,
  return_extremes = FALSE,
  minimum_values_for_solving = 5,
  runn_mean_test_length = 5,
  runn_mean_test_diff = 5,
  daily_patch_max_mean_bias = NA,
  daily_patch_max_stdev_bias = NA
)

Arguments

Details

Many agroclimatic metrics are calculated from hourly temperature data. chillR provides functions for generating hourly data from daily records, which are often available. Small gaps in such daily records can easily be closed through linear interpolation, with relatively small errors, so that complete hourly records can be generated. However, many sites have recorded actual hourly temperatures, which allow much more accurate site-specific assessments. Such records quite often have gaps, which need to be closed before calculating most agroclimatic metrics (such as Chill Portions). Linear interpolation is not a good option for this, because daily temperature curves are not linear. Moreover, when gaps exceed a certain number of hours, important featured would be missed (e.g. interpolating between temperatures at 8 pm and 8 am may miss all the cool hours of the day, which would greatly distort chill estimates).

This function solves this problem by using an idealized daily temperature curve as guide to the interpolation of hourly temperature data.

These are the steps: 1) produce an idealized temperature curve for the site (which requires site latitude as an input), assuming minimum and maximum temperatures of 0 and 1 degrees C, respectively. The calculations are based on equations published by Spencer (1971), Almorox et al. (2005) and Linvill (1990, though I modified these slightly to produce a smooth curve). This curve describes the expected relationship of the temperature for the respective hour with minimum and maximum temperatures of the same, previous or next day (depending on the time of day), according to idealized temperature curve. At this point, however, these daily minimum or maximum temperatures aren't known yet.

2) determine minimum and maximum temperatures for each day. For each minimum and maximum daily temperature, the expected relationships between hourly temperatures and daily extremes determined in step 1, combined with the hourly temperatures that were observed can be interpreted as an overdetermined set of equations that define these temperatures. Since few days will follow the ideal curve precisely, and there are usually more than two equations that define the same daily temperature extreme value, these equations can only be solved numerically. This is implemented with the qr.solve function, which can provide estimates of the minimum and maximum temperatures for all days from the available hourly records.

3) interpolate gaps in the record of estimated daily temperature extremes. There can be days, when the number of recorded hourly temperatures isn't sufficient for inferring daily minimum or maximum temperatures. The resulting gaps are closed by linear interpolation (this may produce poor results if gaps are really large, but this isn't currently addressed).

4) compute an idealized daily temperature curve for all days, based on estimated daily temperature extremes (using the make_hourly_temperatures function).

5) calculate deviation of recorded temperatures from idealized curve.

6) linearly interpolate deviation values using the interpolate_gaps function.

7) add interpolated deviation values to idealized temperature curve.

Value

data frame containing interpolated temperatures for all hours within the interval defined by the first and last day of the hourtemps input.

Author(s)

Eike Luedeling

References

Linvill DE, 1990. Calculating chilling hours and chill units from daily maximum and minimum temperature observations. HortScience 25(1), 14-16.

Spencer JW, 1971. Fourier series representation of the position of the Sun. Search 2(5), 172.

Almorox J, Hontoria C and Benito M, 2005. Statistical validation of daylength definitions for estimation of global solar radiation in Toledo, Spain. Energy Conversion and Management 46(9-10), 1465-1471)

Examples

Winters_gaps<-make_JDay(Winters_hours_gaps[1:2000,])
colnames(Winters_gaps)[5:6]<-c("Temp","original_Temp")
interp<-interpolate_gaps_hourly(hourtemps=Winters_gaps,latitude=38.5)

#plot results: interpolated temperatures are shown in red, measured temperatures in black.
plot(interp$weather$Temp[1:120]~c(interp$weather$JDay[1:120]+
   interp$weather$Hour[1:120]/24),type="l",
   col="RED",lwd=2,xlab="JDay",ylab="Temperature")
lines(interp$weather$Temp_measured[1:120]~c(interp$weather$JDay[1:120]+
   interp$weather$Hour[1:120]/24),lwd=2)

Description

This function counts the days between two Julian dates, taking into account whether the season extends past the end of a calender year and whether the count is to be done for a leap year.

Usage

JDay_count(start_date, end_date, season = NA, leap_year = FALSE)

Arguments

Value

Boolean result (TRUE/FALSE) of the comparison.

Author(s)

Eike Luedeling

Examples

JDay_count(start_date=320,end_date=20,season=c(305,59),leap_year=2004)

Description

For two Julian dates, this function checks whether the first date is earlier than the second date within a user-defined phenological season. This is particularly useful for seasons that start in one year and end in the next, because simple > or < operations can produce wrong results then.

Usage

JDay_earlier(check_date, ref_date, season = c(1, 366))

Arguments

Value

Boolean result (TRUE/FALSE) of the comparison.

Author(s)

Eike Luedeling

Examples

JDay_earlier(check_date=10,ref_date=365,season=c(305,59))

Description

For two Julian dates, this function checks whether the first date is later than the second date within a user-defined phenological season. This is particularly useful for seasons that start in one year and end in the next, because simple > or < operations can produce wrong results then.

Usage

JDay_later(check_date, ref_date, season = c(1, 366))

Arguments

Value

Boolean result (TRUE/FALSE) of the comparison.

Author(s)

Eike Luedeling

Examples

JDay_later(check_date=10,ref_date=365,season=c(305,59))

Description

Bloom data of sweet cherry var. 'Schneiders spaete Knorpelkirsche' recorded at Klein-Altendorf, Germany, the experimental station of the University of Bonn

Format

A data frame with the following 2 variables.

Year

a numeric vector, indicating the observation year

pheno

a vector that, when coerced by as.numeric, contains bloom data in Julian dates (day of the year)

Source

data were collected by Achim Kunz and Michael Blanke, University of Bonn

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

data(KA_bloom)

Description

Daily temperature data from Klein-Altendorf, Germany, for use in combination with the example phenology dataset KA_bloom.

Format

A data frame with observations on the following 5 variables.

Year

a numeric vector - the observation year

Month

a numeric vector - the observation month

Day

a numeric vector - the observation day

Tmax

a numeric vector - daily maximum temperature

Tmin

a numeric vector - daily minimum temperature

Source

data were collected by Achim Kunz and Michael Blanke, University of Bonn

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

data(KA_weather)

Description

This function determines whether a given year is a leap year

Usage

leap_year(x)

Arguments

Details

Takes a year number as input, and returns TRUE if this is a leap year, and FALSE if not

Value

boolean variable (TRUE or FALSE)

Author(s)

Eike Luedeling, but based on pseudocode from Wikipedia

References

https://en.wikipedia.org/wiki/Leap_year

Examples

leap_year(2015)  
leap_year(2016)

Description

This is a slightly modified version of the load_temperature_scenarios function that can load climate scenarios downloaded with the getClimateWizardData and saved with the save_temperature_scenarios function. This separate function is necessary, because the climate scenarios are expressed as lists, with one element being a data.frame.

Usage

load_ClimateWizard_scenarios(path, prefix)

Arguments

Value

a list of temperature scenarios.

Author(s)

Eike Luedeling

Examples

temps<-list(Element1=data.frame(a=1,b=2),Element2=data.frame(a=c(2,3),b=c(8,4)))
# save_temperature_scenarios(temps,path=getwd(),prefix="temperatures")
# temps_reloaded<-load_temperature_scenarios(path=getwd(),prefix="temperatures")

Description

The temperature_generation can produce synthetic temperature scenarios, but it can take a while to run, especially for large ensembles of climate scenarios. The save_temperature_scenarios function can then save these scenarios to disk as a series of .csv files, so that they can later be used again, without re-running the generation function. Conversely, the load_temperature_scenarios function allows reading the data back into R. This function also works with any other list of data.frames.

Usage

load_temperature_scenarios(path, prefix)

Arguments

Value

a list of temperature scenarios.

Author(s)

Eike Luedeling

Examples

temps<-list(Element1=data.frame(a=1,b=2),Element2=data.frame(a=c(2,3),b=c(8,4)))
# save_temperature_scenarios(temps,path=getwd(),prefix="temperatures")
# temps_reloaded<-load_temperature_scenarios(path=getwd(),prefix="temperatures")

Description

Time series often have gaps, and these are often not marked by 'no data' values but simply missing from the dataset. This function completes the time series by adding lines for all these missing records. For these lines, all values are set to 'NA'. By setting timestep<-"hour", this function can also process hourly data. Where data are provided at a time resolution that is finer than timestep, values are aggregated (by calculating the mean) to timestep resolution (e.g. when data are at 15-minute resolution, they will be aggregated to hourly average values - at timestep=="hour" - or daily average values - at timestep=="day").

Usage

make_all_day_table(
  tab,
  timestep = "day",
  input_timestep = timestep,
  tz = "GMT",
  add.DATE = TRUE,
  no_variable_check = FALSE,
  aggregation_hours = NULL
)

Arguments

Value

data frame containing all the columns of the input data frame, but one row for each day between the start and end of the dataset. Data values for the missing rows are filled in as 'NA'. Dates are expressed as c("YEARMODA","DATE","Year","Month","Day"). In this, 'DATE' is the date in ISOdate format.

Author(s)

Eike Luedeling

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

#fill in missing lines in a weather dataset (modified from KA_weather)
day_to_day<-make_all_day_table(KA_weather[c(1:10,20:30),],timestep="day")

#fill in missing hours in the Winters_hours_gaps dataset
Winters_hours<-subset(Winters_hours_gaps, select = -c(Temp_gaps))[1:2000,]
hour_to_hour<-make_all_day_table(Winters_hours,timestep="hour",input_timestep="hour")

#convert Winters_hours_gaps dataset into daily temperature data (min, max, mean)
hour_to_day<-make_all_day_table(Winters_hours,timestep="day",input_timestep="hour")
hour_to_day<-make_all_day_table(Winters_hours,timestep="day",input_timestep="hour",
                               aggregation_hours=c(3,3,2))

Description

Makes a list of the weather stations contained in the UC IPM database, with geographic coordinates. This requires parsing through quite a few websites, because the coordinates don't seem to be stored in one central (and easily accessible) place. Hence this is much slower than one might expect. A shortcut is the california_stations dataset supplied with chillR, which contains the result of running this function in February 2016. The default in the other relevant functions will be the use of this pre-stored list, but if the current station coverage is needed, this function can help. Having said this, station coverage probably won't change very rapidly, so in most cases, the california_stations dataset should be enough.

Usage

make_california_UCIPM_station_list()

Value

a data.frame containing stations from the California UC IPM database (), with the following columns: "Name", "Code", "Interval", "Lat", "Long", "Elev".

Author(s)

Eike Luedeling

References

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

#cali_stats<-make_california_UCIPM_station_list()

Description

This function generates a plot of a climate metric over multiple years, including an indication of data quality, i.e. the share of missing values. Output can be either an R plot or a .png image

Usage

make_chill_plot(
  chill,
  model,
  start_year = NA,
  end_year = NA,
  metriclabel = NULL,
  yearlabel = "End_year",
  misstolerance = 10,
  image_type = NA,
  outpath = NA,
  filename = NA,
  fonttype = "serif",
  plotylim = NA,
  plottitle = NULL
)

Arguments

Details

Plots climatic metrics computed with chilling or tempResponse, indicating the completeness of the temperature record by shades of gray.

Value

only a side effect - plot of climate metric over time; bars are color coded according to the number of missing values. Bars with numbers of missing values above the misstolerance are not show and instead marked '*' (to distinguish them from 0 counts)

Author(s)

Eike Luedeling

Examples

make_chill_plot(tempResponse(stack_hourly_temps(fix_weather(KA_weather[KA_weather$Year>2005,]))),
"Chill_Portions",start_year=1990,end_year=2010,metriclabel="Chill Portions")

Description

Function to make climate scenarios for plotting from a list of climate metric data, e.g. produced by tempResponse_daily_list.

Usage

make_climate_scenario(
  metric_summary,
  caption = NULL,
  labels = names(metric_summary),
  time_series = FALSE,
  historic_data = NULL,
  add_to = NULL
)

Arguments

Value

a list of climate scenario objects, which can be supplied to plot_climate_scenarios.

Author(s)

Eike Luedeling

Examples

chill<-chilling(stack_hourly_temps(fix_weather(KA_weather[which(KA_weather$Year>1990),]),
   latitude=50.4))
multi_chills<-list('2001'=chill,'2005'=chill,'2009'=chill)
chills_to_plot<-make_climate_scenario(multi_chills,caption=c("Historic","data"),
   time_series=TRUE,historic_data=chill)
chills_to_plot<-make_climate_scenario(multi_chills,caption=c("Future1"),add_to=chills_to_plot)
chills_to_plot<-make_climate_scenario(multi_chills,caption=c("Future2"),add_to=chills_to_plot)
plot_climate_scenarios(chills_to_plot,metric="Chill_portions",metric_label="Chill Portions")

Description

Many climate scenarios we may want to plot consist of data stored across many files. These files typically contain certain character strings that mark, e.g. the RCP scenario or the point in time. This function facilitates accessing such files by allowing the specification of search string (criteria_list), according to which files are selected. They are then converted into climate_scenario files that can become part of a list passed to plot_climate_scenarios for plotting.

Usage

make_climate_scenario_from_files(
  metric_folder,
  criteria_list,
  caption = NULL,
  time_series = FALSE,
  labels = NULL,
  historic_data = NULL
)

Arguments

Value

a climate scenario object, which can be part of a list supplied to plot_climate_scenarios.

Author(s)

Eike Luedeling

Examples

# historic_scenario<-make_climate_scenario(metric_folder=chillout_folder,
#                                          criteria_list=list(cult,c(1975,2000,2015)),
#                                          caption=c("Historic","data"),
#                                         time_series=TRUE,
#                                         labels=c(1975,2000,2015),
#                                         historic_data=historic_data)

Description

Function to make figures showing the mean rate of chill and heat accumulation for each day of the year, as well as as the standard deviation.

Usage

make_daily_chill_figures(
  daily_chill,
  file_path,
  models = c("Chilling_Hours", "Utah_Chill_Units", "Chill_Portions", "GDH"),
  labels = NA
)

Arguments

Details

Chill metrics are calculated as given in the references below. Chilling Hours are all hours with temperatures between 0 and 7.2 degrees C. Units of the Utah Model are calculated as suggested by Richardson et al. (1974) (different weights for different temperature ranges, and negation of chilling by warm temperatures). Chill Portions are calculated according to Fishman et al. (1987a,b). More honestly, they are calculated according to an Excel sheet produced by Amnon Erez and colleagues, which converts the complex equations in the Fishman papers into relatively simple Excel functions. These were translated into R. References to papers that include the full functions are given below. Growing Degree Hours are calculated according to Anderson et al. (1986), using the default values they suggest. This function uses the Kendall package.

Value

data frame containing all information used to make the figures that are saved. For each Julian Date, means and standard deviations of all chill and heat metrics are saved. In addition, Mann-Kendall tests are performed for daily accumulations of all metrics. p and tau values from this test indicate the level of statistical significance. This non-parametric test is reliable for time series data.

Note

After doing extensive model comparisons, and reviewing a lot of relevant literature, I do not recommend using the Chilling Hours or Utah Models, especially in warm climates! The Dynamic Model (Chill Portions), though far from perfect, seems much more reliable.

Author(s)

Eike Luedeling

References

Model references:

Chilling Hours:

Weinberger JH (1950) Chilling requirements of peach varieties. Proc Am Soc Hortic Sci 56, 122-128

Bennett JP (1949) Temperature and bud rest period. Calif Agric 3 (11), 9+12

Utah Model:

Richardson EA, Seeley SD, Walker DR (1974) A model for estimating the completion of rest for Redhaven and Elberta peach trees. HortScience 9(4), 331-332

Dynamic Model:

Erez A, Fishman S, Linsley-Noakes GC, Allan P (1990) The dynamic model for rest completion in peach buds. Acta Hortic 276, 165-174

Fishman S, Erez A, Couvillon GA (1987a) The temperature dependence of dormancy breaking in plants - computer simulation of processes studied under controlled temperatures. J Theor Biol 126(3), 309-321

Fishman S, Erez A, Couvillon GA (1987b) The temperature dependence of dormancy breaking in plants - mathematical analysis of a two-step model involving a cooperative transition. J Theor Biol 124(4), 473-483

Growing Degree Hours:

Anderson JL, Richardson EA, Kesner CD (1986) Validation of chill unit and flower bud phenology models for 'Montmorency' sour cherry. Acta Hortic 184, 71-78

Model comparisons and model equations:

Luedeling E, Zhang M, Luedeling V and Girvetz EH, 2009. Sensitivity of winter chill models for fruit and nut trees to climatic changes expected in California's Central Valley. Agriculture, Ecosystems and Environment 133, 23-31

Luedeling E, Zhang M, McGranahan G and Leslie C, 2009. Validation of winter chill models using historic records of walnut phenology. Agricultural and Forest Meteorology 149, 1854-1864

Luedeling E and Brown PH, 2011. A global analysis of the comparability of winter chill models for fruit and nut trees. International Journal of Biometeorology 55, 411-421

Luedeling E, Kunz A and Blanke M, 2011. Mehr Chilling fuer Obstbaeume in waermeren Wintern? (More winter chill for fruit trees in warmer winters?). Erwerbs-Obstbau 53, 145-155

Review on chilling models in a climate change context:

Luedeling E, 2012. Climate change impacts on winter chill for temperate fruit and nut production: a review. Scientia Horticulturae 144, 218-229

The PLS method is described here:

Luedeling E and Gassner A, 2012. Partial Least Squares Regression for analyzing walnut phenology in California. Agricultural and Forest Meteorology 158, 43-52.

Wold S (1995) PLS for multivariate linear modeling. In: van der Waterbeemd H (ed) Chemometric methods in molecular design: methods and principles in medicinal chemistry, vol 2. Chemie, Weinheim, pp 195-218.

Wold S, Sjostrom M, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab 58(2), 109-130.

Mevik B-H, Wehrens R, Liland KH (2011) PLS: Partial Least Squares and Principal Component Regression. R package version 2.3-0. http://CRAN.R-project.org/package0pls.

Some applications of the PLS procedure:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Yu H, Luedeling E and Xu J, 2010. Stronger winter than spring warming delays spring phenology on the Tibetan Plateau. Proceedings of the National Academy of Sciences (PNAS) 107 (51), 22151-22156.

Yu H, Xu J, Okuto E and Luedeling E, 2012. Seasonal Response of Grasslands to Climate Change on the Tibetan Plateau. PLoS ONE 7(11), e49230.

The exact procedure was used here:

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

The chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

weather<-fix_weather(KA_weather[which(KA_weather$Year>2005),])

dc<-daily_chill(stack_hourly_temps(weather,50.4), 11,models=list(Chill_Portions=Dynamic_Model))

# md<-make_daily_chill_figures(dc, paste(getwd(),"/daily_chill_",sep=""),models="Chill_Portions",
#  labels="Chill Portions")

Description

This function generates a plot of the accumulation of a climate metric throughout the year. Its standard output are the mean daily accumulation and the standard deviation. It is also possible to add one or several so-called focusyears to add the daily accumulation during these years to the plots. Plots can be produced in R or directly exported as .png files.

Usage

make_daily_chill_plot(
  daily_chill,
  metrics = NA,
  startdate = 1,
  enddate = 366,
  useyears = NA,
  metriclabels = NA,
  focusyears = "none",
  cumulative = FALSE,
  image_type = NA,
  outpath = NA,
  filename = NA,
  fonttype = "serif",
  title = NA,
  plotylim = NA
)

Arguments

Details

Plots daily accumulation of climatic metrics, such as winter chill, as daily accumulation rates or as cumulative accumulation. A legend is only added, when focusyears are also shown. Otherwise the plot is reasonably self-explanatory.

Value

The main purpose of the function is a side effect - plots of daily climate metric accumulation. However, all the data used for making the plots is returned as a list containing an element for each metric, which consists of a data.table with the daily means, standard deviation and daily values for all focusyears.

Author(s)

Eike Luedeling

Examples

day_chill<-make_daily_chill_plot(daily_chill(stack_hourly_temps(fix_weather(
  KA_weather[which(KA_weather$Year>2005),])),
  running_mean=11),focusyears=c(2001,2005),cumulative=TRUE,startdate=300,enddate=30)

Description

This function generates a plot of the accumulation of a climate metric throughout the year. Its standard output are the mean daily accumulation and the standard deviation. It is also possible to add one or several so-called focusyears to add the daily accumulation during these years to the plots. Plots can be produced in R or directly exported as .png files.

Usage

make_daily_chill_plot2(
  daily,
  metrics = NA,
  startdate = 1,
  enddate = 366,
  useyears = NA,
  metriclabels = NA,
  focusyears = "none",
  cumulative = FALSE,
  fix_leap = TRUE
)

Arguments

Details

Plots daily accumulation of climatic metrics, such as winter chill, as daily accumulation rates or as cumulative accumulation. A legend is only added, when focusyears are also shown. Otherwise the plot is reasonably self-explanatory.

Value

The main purpose of the function is a side effect - plots of daily climate metric accumulation. However, all the data used for making the plots is returned as a list containing an element for each metric, which consists of a data.table with the daily means, standard deviation and daily values for all focusyears.

Author(s)

Eike Luedeling

Examples

daily<-daily_chill(stack_hourly_temps(fix_weather(
  KA_weather[which(KA_weather$Year>2005),])),running_mean=11)

make_daily_chill_plot2(daily,metrics=c("Chill_Portions","GDH"),cumulative=TRUE,
   startdate=300,enddate=30,focusyears=c(2009,2008))

Description

This function generates hourly temperature records for a particular location from daily minimum and maximum temperatures and latitude.

Usage

make_hourly_temps(latitude, year_file, keep_sunrise_sunset = FALSE)

Arguments

Details

Temperature estimates are based on an idealized daily temperature curve that uses a sine curve for daytime warming and a logarithmic decay function for nighttime cooling. The input data frame can have more columns, which are preserved, but ignored in the processing. References to papers outlining the procedures are given below.

Note that this function should be able to generate hourly temperatures for all latitudes, but it uses an algorithm designed for locations with regular day/night behavior. It may therefore be that the curves aren't very realistic for very short or very long days, or especially for polar days and nights.

Value

data frame containing all the columns of year_file, plus 24 columns for hourly temperatures (called Hour_1 ... Hour_24).

Author(s)

Eike Luedeling

References

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Luedeling E, Girvetz EH, Semenov MA and Brown PH, 2011. Climate change affects winter chill for temperate fruit and nut trees. PLoS ONE 6(5), e20155.

The temperature interpolation is described in

Linvill DE, 1990. Calculating chilling hours and chill units from daily maximum and minimum temperature observations. HortScience 25(1), 14-16.

Calculation of sunrise, sunset and daylength was done according to

Spencer JW, 1971. Fourier series representation of the position of the Sun. Search 2(5), 172.

Almorox J, Hontoria C and Benito M, 2005. Statistical validation of daylength definitions for estimation of global solar radiation in Toledo, Spain. Energy Conversion and Management 46(9-10), 1465-1471)

Examples

weather<-fix_weather(KA_weather)

THourly<-make_hourly_temps(50.4,weather$weather)

#in most cases, you're probably better served by stack_hour_temperatures

Description

This function produced Julian Dates (days of the year) from columns "Day", "Month" and "Year" in a dataframe.

Usage

make_JDay(dateframe)

Arguments

Value

Returns the same data.frame, but with column "JDay" added. This then contains the Julian Dates.

Author(s)

Eike Luedeling

References

The chillR package:

Examples

dates<-data.frame(Year=c(1977,1980,2004,2011,2016),Month=c(11,8,3,12,8),Day=c(1,21,2,24,2)) 
make_JDay(dates)

Description

For multiple datasets, this function plots surface plots relating mean temperatures during specified periods to annually recurring variables (e.g. flowering). It produces one panel per dataset and plots them all in one figure. Plots can be produced in R or directly exported as .png files.

Usage

make_multi_pheno_trend_plot(
  pheno_list,
  fixed_weather,
  split_month = 6,
  outpath = NA,
  file_name = NA,
  image_type = "png",
  fonttype = "serif",
  percol = 5,
  xlabel = NA,
  ylabel = NA,
  height_factor = 0.8
)

Arguments

Details

This function is only useful, if you want to plot several surface plots in the same figure. These must relate to the same weather dataset. Arguably, this function isn't quite ready to be released, but it performs some useful functions that you may be interested in...

Value

Only a side effect is produced: either a .png file or an R graphic showing the multi-panel contour figure.

Author(s)

Eike Luedeling

Examples

#this example uses arbitrarily modified versions of the KA_bloom dataset, and the starts
#end ends of the periods are also arbitraty. So the outputs may not make a lot of sense...

weather<-fix_weather(KA_weather[which(KA_weather$Year>2000),])
pheno_list<-data.frame(varieties=c("KA1","KA2","KA3","KA4"), Start_chill=c(270,305,315,320),
  End_chill=c(15,20,35,40), Start_heat=c(17,25,40,45),End_heat=c(90,100,110,115),
  Link=c("KA1.csv","KA2.csv","KA3.csv","KA4.csv"))
  
 # write.csv(KA_bloom,"KA1.csv",row.names=FALSE)
 KA_bloom$pheno<-as.numeric(as.character(KA_bloom$pheno))+10
 # write.csv(KA_bloom,"KA2.csv",row.names=FALSE)
 KA_bloom$pheno<-KA_bloom$pheno+10
 # write.csv(KA_bloom,"KA3.csv",row.names=FALSE)
 KA_bloom$pheno<-KA_bloom$pheno+10
 # write.csv(KA_bloom,"KA4.csv",row.names=FALSE)

# make_multi_pheno_trend_plot(pheno_list,weather, split_month=6,
#            outpath=NA,file_name=NA,image_type="",fonttype="serif",percol=2)

Description

The timing of many developmental stages of temperate plants is understood to depend on temperatures during two phases. This function seeks to illustrate this dependency by plotting phenological dates as colored surface, as a function of mean temperatures during both phases, which are indicated on the x and y axes.

Usage

make_pheno_trend_plot(
  weather_data_frame,
  split_month = 6,
  pheno,
  use_Tmean = FALSE,
  Start_JDay_chill,
  End_JDay_chill,
  Start_JDay_heat,
  End_JDay_heat,
  outpath,
  file_name,
  plot_title,
  ylabel = NA,
  xlabel = NA,
  legend_label = NA,
  image_type = "png",
  colorscheme = "normal",
  fonttype = "serif"
)

Arguments

Details

The generation of the color surface is based on the Kriging technique, which is typically used for interpolation of spatial data. The use for this particular purpose is a bit experimental. The function uses the Krig function from the fields package.

Value

Author(s)

Eike Luedeling

References

Guo L, Dai J, Wang M, Xu J, Luedeling E, 2015. Responses of spring phenology in temperate zone trees to climate warming: a case study of apricot flowering in China. Agricultural and Forest Meteorology 201, 1-7.

Guo L, Dai J, Ranjitkar S, Xu J, Luedeling E, 2013. Response of chestnut phenology in China to climate variation and change. Agricultural and Forest Meteorology 180, 164-172.

Luedeling E, Guo L, Dai J, Leslie C, Blanke M, 2013. Differential responses of trees to temperature variation during the chilling and forcing phases. Agricultural and Forest Meteorology 181, 33-42.

the interpolation was done according to:

Furrer, R., Nychka, D. and Sain, S., 2012. Fields: Tools for spatial data. R package version 6.7.

Reference to the chillR package:

Luedeling E, Kunz A and Blanke M, 2013. Identification of chilling and heat requirements of cherry trees - a statistical approach. International Journal of Biometeorology 57,679-689.

Examples

weather<-fix_weather(KA_weather)

#the output of the PLS function (PLS_pheno, plotted with plot_PLS) can be used to select
#phases that are likely relevant for plant phase timing. See respective examples for running
#these functions.

file_path<-paste(getwd(),"/",sep="")

#mpt<-make_pheno_trend_plot(weather_data_frame = weather$weather, split_month = 6,
#      pheno = KA_bloom, use_Tmean = FALSE, Start_JDay_chill = 260, 
#      End_JDay_chill = 64, Start_JDay_heat = 44, End_JDay_heat = 103,
#      outpath = file_path, file_name = "pheno_trend_plot",
#      plot_title = "Impacts of chilling and forcing temperatures on cherry phenology",
#      ylabel = NA, xlabel = NA, legend_label = NA, image_type = "png", 
#      colorscheme = "normal")