Merge pull request #51 from tcarleton/50-fix_projection

50 fix projection and 49 update documentation

Author: tliddell4

R/secondary_weights.R

@@ -29,6 +29,16 @@
 #' @export
 secondary_weights <- function(secondary_raster, grid = era5_grid, extent = "full"){
 
+  ## check to make sure climate raster is a spatraster, change if not
+  if (!inherits(grid, "SpatRaster")) {
+    clim_raster <- terra::rast(grid)
+  } else {
+
+    clim_raster <- grid
+
+  }
+
+
   ## Return error if terra::extent can't inherit from the value supplied
   ## won't be able to check if secondary raster fully overlaps if
   ## this input isn't compatible with terra::extent()
@@ -72,14 +82,6 @@ secondary_weights <- function(secondary_raster, grid = era5_grid, extent = "full
 
   }
 
-  ## check to make sure climate raster is a spatraster, change if not
-  if (!inherits(grid, "SpatRaster")) {
-    clim_raster <- terra::rast(grid)
-  } else {
-
-    clim_raster <- grid
-
-  }
 
   ## check if SpatRaster is in geographic coodrinates
   if(!terra::is.lonlat(clim_raster)) {
@@ -131,7 +133,8 @@ secondary_weights <- function(secondary_raster, grid = era5_grid, extent = "full
 
       extent <- extent
 
-    }
+  }
+
 
   ## If an extent was included, crop it to the extent to save ram
   ## -----------------------------------------------
@@ -208,6 +211,15 @@ secondary_weights <- function(secondary_raster, grid = era5_grid, extent = "full
 
   }
 
+  # Reproject if necessary
+  if(terra::crs(secondary_raster, TRUE) != terra::crs(grid, TRUE)){
+
+    message(crayon::yellow("Warning: reprojecting secondary_raster to match grid, which may affect data"))
+    secondary_raster <- terra::project(secondary_raster, clim_raster)
+
+  }
+
+
 
   ## crop the ERA/climate raster to the appropriate extent
   ## use the extent of the previously user-cropped secondary raster
@@ -216,9 +228,6 @@ secondary_weights <- function(secondary_raster, grid = era5_grid, extent = "full
   # Find the difference between the climate raster resolution and secondary raster resolution
   clim_raster <- terra::crop(clim_raster, terra::ext(secondary_raster), snap="out")
 
-  ## set crs of secondary raster to match climate data
-  ## -----------------------------------------------
-  terra::crs(secondary_raster) <- terra::crs(clim_raster)
 
   ## check if the cropped secondary raster contains NA values
   if(isTRUE(any(is.na(terra::values(secondary_raster, na.rm=FALSE))))) {

R/staggregate.R

@@ -431,14 +431,14 @@ polygon_aggregation <- function(clim_dt, weights_dt, list_names, time_agg, weigh
 #'   unless `daily_agg = 'none'`)
 #' @param start_date the date (and time, if applicable) of the first layer in
 #'  the stack. To be input in a format compatible with
-#'  lubridate::as_datetime(), e.g. `"1991-10-29"` or `"1991-10-29 00:00:00"`.
+#'  lubridate::as_datetime(), e.g. `'1991-10-29'` or `'1991-10-29 00:00:00'`.
 #'  The default is `NA` since the spatRasters usually already contain temporal
 #'  information in the layer names and they do not need to be manually supplied.
 #' @param time_interval the time interval between layers in the spatRaster to be
-#'  aggregated. To be input in a format compatible with seq(), e.g.
-#'  `'1 day'` or `'3 months'`. The default is `'1 hour'` and this argument is
-#'  required if daily_agg is not `'none'` or if the `start_date` argument is not
-#'  `NA`.
+#'  aggregated. To be input in a format compatible with
+#'  `lubridate::as.duration()` such as `'30 minutes'`, `'2 hours'`, `'1 day'`,
+#'  etc. The default is `'1 hour'` and this argument is required if daily_agg is
+#'  not `'none'` or if the `start_date` argument is not `NA`.
 #' @param weights_join_tolerance the tolerance to use when joining
 #' overlay_weights with the climate data by the x and y columns. This is useful
 #' when the height/width of your data cells expressed in degrees is a very long
@@ -588,14 +588,14 @@ staggregate_polynomial <- function(data, overlay_weights, daily_agg, time_agg =
 #'   unless `daily_agg = 'none'`)
 #' @param start_date the date (and time, if applicable) of the first layer in
 #'  the stack. To be input in a format compatible with
-#'  lubridate::as_datetime(), e.g. `"1991-10-29"` or `"1991-10-29 00:00:00"`.
+#'  lubridate::as_datetime(), e.g. `'1991-10-29'` or `'1991-10-29 00:00:00'`.
 #'  The default is `NA` since the spatRasters usually already contain temporal
 #'  information in the layer names and they do not need to be manually supplied.
 #' @param time_interval the time interval between layers in the spatRaster to be
-#'  aggregated. To be input in a format compatible with seq(), e.g.
-#'  `'1 day'` or `'3 months'`. The default is `'1 hour'` and this argument is
-#'  required if daily_agg is not `'none'` or if the `start_date` argument is not
-#'  `NA`.
+#'  aggregated. To be input in a format compatible with
+#'  `lubridate::as.duration()` such as `'30 minutes'`, `'2 hours'`, `'1 day'`,
+#'  etc. The default is `'1 hour'` and this argument is required if daily_agg is
+#'  not `'none'` or if the `start_date` argument is not `NA`.
 #' @param weights_join_tolerance the tolerance to use when joining
 #' overlay_weights with the climate data by the x and y columns. This is useful
 #' when the height/width of your data cells expressed in degrees is a very long
@@ -773,14 +773,14 @@ staggregate_spline <- function(data, overlay_weights, daily_agg, time_agg = "mon
 #'   unless `daily_agg = 'none'`)
 #' @param start_date the date (and time, if applicable) of the first layer in
 #'  the raster. To be input in a format compatible with
-#'  lubridate::as_datetime(), e.g. `"1991-10-29"` or `"1991-10-29 00:00:00"`.
+#'  lubridate::as_datetime(), e.g. `'1991-10-29'` or `'1991-10-29 00:00:00'`.
 #'  The default is `NA` since the rasters usually already contain temporal
 #'  information in the layer names and they do not need to be manually supplied.
-#' @param time_interval the time interval between layers in the raster to be
-#'  aggregated. To be input in a format compatible with seq(), e.g.
-#'  `'1 day'` or `'3 months'`. The default is `'1 hour'` and this argument is
-#'  required if daily_agg is not `'none'` or if the `start_date` argument is not
-#'  `NA`.
+#' @param time_interval the time interval between layers in the spatRaster to be
+#'  aggregated. To be input in a format compatible with
+#'  `lubridate::as.duration()` such as `'30 minutes'`, `'2 hours'`, `'1 day'`,
+#'  etc. The default is `'1 hour'` and this argument is required if daily_agg is
+#'  not `'none'` or if the `start_date` argument is not `NA`.
 #' @param weights_join_tolerance the tolerance to use when joining
 #' overlay_weights with the climate data by the x and y columns. This is useful
 #' when the height/width of your data cells expressed in degrees is a very long
@@ -958,13 +958,14 @@ staggregate_bin <- function(data, overlay_weights, daily_agg, time_agg = "month"
 #'   `'month'`, or `'year'`
 #' @param start_date the date (and time, if applicable) of the first layer in
 #'  the raster. To be input in a format compatible with
-#'  lubridate::as_datetime(), e.g. `"1991-10-29"` or `"1991-10-29 00:00:00"`.
+#'  lubridate::as_datetime(), e.g. `'1991-10-29'` or `'1991-10-29 00:00:00'`.
 #'  The default is `NA` since the rasters usually already contain temporal
 #'  information in the layer names and they do not need to be manually supplied.
-#' @param time_interval the time interval between layers in the raster to be
-#'  aggregated. To be input in a format compatible with seq(), e.g.
-#'  `'1 day'` or `'3 months'`. The default is `'1 hour'` and this argument is
-#'  required if the `start_date` argument is not `NA`.
+#' @param time_interval the time interval between layers in the spatRaster to be
+#'  aggregated. To be input in a format compatible with
+#'  `lubridate::as.duration()` such as `'30 minutes'`, `'2 hours'`, `'1 day'`,
+#'  etc. The default is `'1 hour'` and this argument is required if daily_agg is
+#'  not `'none'` or if the `start_date` argument is not `NA`.
 #' @param weights_join_tolerance the tolerance to use when joining
 #' overlay_weights with the climate data by the x and y columns. This is useful
 #' when the height/width of your data cells expressed in degrees is a very long

README.Rmd

@@ -153,10 +153,8 @@ polynomial_output <- staggregate_polynomial(
   
   data = temp_nj_jun_2024_era5 - 273.15, # A raster brick of our primary data, 
                                          # typically but not necessarily climate 
-                                         # data. For now, data must start at 
-                                         # midnight and be hourly. We're 
-                                         # converting from Kelvin to Celsius 
-                                         # here.
+                                         # data. We're converting from Kelvin to 
+                                         # Celsius here.
   
   overlay_weights = county_weights, # Output from Step 2, determined here by 
                                     # area-normalized cropland weights for grid 
@@ -197,10 +195,8 @@ spline_output <- staggregate_spline(
   
   data = temp_nj_jun_2024_era5 - 273.15, # A raster brick of our primary data, 
                                          # typically but not necessarily climate 
-                                         # data. For now, data must start at 
-                                         # midnight and be hourly. We're
-                                         # converting from Kelvin to Celsius 
-                                         # here.
+                                         # data. We're converting from Kelvin to 
+                                         # Celsius here.
   
   overlay_weights = county_weights, # Output from Step 2, determined here by 
                                     # area-normalized cropland weights for grid
@@ -235,10 +231,8 @@ bin_output <- staggregate_bin(
   
   data = temp_nj_jun_2024_era5 - 273.15, # A raster brick of our primary data, 
                                          # typically but not necessarily climate 
-                                         # data. For now, data must start at 
-                                         # midnight and be hourly. We're 
-                                         # converting from Kelvin to Celsius
-                                         # here.
+                                         # data. We're converting from Kelvin to 
+                                         # Celsius here.
   
   overlay_weights = county_weights,  # Output from Step 2, determined here by 
                                      # area-normalized cropland weights for grid 
@@ -272,10 +266,8 @@ The final transformation offered is degree days, which measures the degrees over
 staggregate_degree_days(
   data = temp_nj_jun_2024_era5 - 273.15, # A raster brick of our primary data, 
                                          # typically but not necessarily climate 
-                                         # data. For now, data must start at 
-                                         # midnight and be hourly. We're
-                                         # converting from Kelvin to Celsius
-                                         # here. 
+                                         # data. We're converting from Kelvin to 
+                                         # Celsius here. 
   
   overlay_weights = county_weights, # Output from Step 2, determined here by 
                                     # area-normalized cropland weights for grid