climateR

ClimateR looks to simplfy the steps needed to get raster climate data in R for Areas of Interest (AOI). It currently provides, or will provide, access to gridded climate sources as well as observation networks. Efforts have been made to reduce parameter variations (eg prcp and ppt) so that all calls require the same inputs.

Number	Dataset	Description	Dates
1	GridMET	Gridded Meteorological Data.	1979-Yesterday
2	Daymet	Daily Surface Weather and Climatological Summaries	1980-2017.
3	TopoWX	Topoclimatic Daily Air Temperature Dataset	1948-2016.
4	PRISM	Parameter-elevation Regressions on Independent Slopes	1981-(Yesterday-1)
5	MACA	Multivariate Adaptive Constructed Analogs	1950-2099
6	LOCA	Localized Constructed Analogs	1950-2100
7	BCCA	Bias Corrected Constructed Analogs	1950-2100
8	NCEP	National Centers for Ecological Prediction.	In progress
9	SARRD	Statistical Asynchronous Regional Regression	1960-2100
10	BCSD	Bias Corrected Spatially Downscaled VIC: Monthly Hydrology	1950-2099
11	OSSEB	Operational Simplified Surface Energy Balance	2000-2015
12	TerraClimate	TerraClimate Monthly Gridded Data	1958-2017
13	CHIRPS	Climate Hazards Group InfraRed Precipitation with Station	1980-Current year
14	EDDI	Evaporative Demand Drought Index	1980-Current year

Installation

install.packages("devtools")
devtools::install_github("mikejohnson51/climateR")

Examples

The climateR package works off the AOI framework established in the AOI R package. This framework is also used in HydroData, and NWM.

To get a climate product, an area of interest must be defined:

AOI = getAOI(state = "CA")

Here we are loading a spatial polygon for the state of California. More examples of contruting AOI calls can be found here.

With an AOI, we can construct a call to a dataset for a parameter(s) and date(s) of choice. Here we are querting the PRISM dataset for maximum and mimimun temperture estimates for October 29, 2018:

system.time({
p = getPRISM(AOI, param = c('tmax','tmin'), startDate = "2018-10-29")
})

>  user  system elapsed 
  0.057   1.054   4.845 

r = raster::stack(p$tmax, p$tmin)
names(r) = c('tmax', 'tmin')
rasterVis::levelplot(r)

Some sources are downscaled Global Climate Models (GCMs). These allow you to query future forecasted ensemble members. One example is from the MACA dataset:

system.time({
m = getMACA(AOI, 
            model = "CCSM4", 
            param = 'prcp', 
            scenario = c('rcp45', 'rcp85'), 
            startDate = "2080-06-29", 
            endDate = "2080-06-30")
})

>  user  system elapsed 
  0.183   1.213   5.998

r = raster::stack(m$prcp_rcp45, m$prcp_rcp85)
names(r) = paste0(c(rep("RCP45_", 2), rep("RCP85_", 2)), unique(substring(names(r), 2, 11)))
rasterVis::levelplot(r, par.settings = rasterTheme(region=sequential_hcl(10, power=2.2)))

Large scale data grabs are also quite efficient

system.time({
  g = getAOI(state = "conus") %>%  getGridMET(param = 'srad', startDate = "2017-06-29")
})

>   user  system elapsed 
  0.304   1.618   6.063 

raster::plot(g$srad, col = viridis::viridis(100), axes = F, box= F)
title(main = "Solar Radiation 2017-06-29\n4km Resolution")
sp::plot(g$AOI, add = T)

Getting ensemble averages is also quite simple:

models = c('ccsm', 'cnrm', 'csiro', 'hadgem', 'pcm')

system.time({
  temp = getAOI(state = "conus") %>% getSARRD(param = 'tmin', model = models, ensemble = 'a2', startDate = "2080-11-29")
})

> user  system elapsed 
0.554   1.798  36.724

#stack model outputs
s = stack(temp)

# add mean to stack
s = addLayer(s, mean(s))
names(s) = c(models, "Ensemble")

# Plot
levelplot(s, par.settings = RdBuTheme)

Statistics are not limited to mean:

stats= stack(max(s), min(s), mean(s), max(s) - min(s), calc(s, sd), sum((s - mean(s))^2) / 5)
names(stats) = c("Ensemble Max", "Ensemble Min", "Ensemble Mean",  "Ensemble Range",  "Ensemble SD",  "Ensemble Variance")

Not all datasets are USA focused. The TerraClimate Dataset offers global, monthly data up to the current year. Here we can get the PDSI (Palmer Index for October 2017):

w = AOI::world %>% getTerraClim(param = 'palmer', startDate = '2017-10-10')

Finally, data gathering is not limitied to areas and can be done as a time series for a point. Here we look for a 10-model ensemble for maximum temperture in Colorado Springs, Colorado from the LOCA downscaled dataset between Jan 1, 2030 and Decemeber 31, 2040.

system.time({
  ts = geocode('Colorado Springs') %>% 
       getLOCA(param = 'tmax', model = 10, startDate = "2030-01-01", endDate = "2040-12-31")
})
> user  system elapsed 
2.716   4.506  193.675 

In a little over 3 minutes we collected 40,180 data points (~208 per second). The results can be seen below:

Support:

ClimateR is written by Mike Johnson, a graduate Student at the University of California, Santa Barbara in Keith C. Clarke's Lab, 2018 and is funded through the NOAA National Water Center (NWC) via the UCAR COMET Program (2017/18).