add CalAdapt WRF download function for CMIP6 hourly met data #3
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,260 @@ | ||
| #' Download Cal-Adapt WRF CMIP6 outputs for a single site and convert to CF | ||
| #' | ||
| #' Fetches hourly WRF dynamically downscaled data from the Cal-Adapt Analytics | ||
| #' Engine (CADCAT S3 bucket) via caladaptR, extracts the nearest grid cell to | ||
| #' the site, converts units to CF-1.8, and writes one NetCDF per year. | ||
| #' | ||
| #' WRF grids are cached in tempdir() so that when met.process calls this for | ||
| #' multiple sites in the same R session, each grid is only fetched from S3 once. | ||
| #' For 200 sites x 8 vars x 20 years that cuts S3 round trips from 32,000 to 160. | ||
| #' | ||
| #' Available models (all at 45 km, ssp370): CESM2, CNRM-ESM2-1, EC-Earth3, | ||
| #' EC-Earth3-Veg, FGOALS-g3, MPI-ESM1-2-HR, MIROC6, TaiESM1. | ||
| #' Only CESM2 has ssp245 and ssp585 in addition to ssp370. | ||
| #' | ||
| #' Precipitation for MPI-ESM1-2-HR, MIROC6, and TaiESM1 is derived from | ||
| #' rainc + rainnc components; caladaptR handles this transparently. | ||
| #' | ||
| #' @param outfolder Directory for storing output | ||
| #' @param start_date Start date for met data | ||
| #' @param end_date End date for met data | ||
| #' @param site_id BETY site id | ||
| #' @param lat.in Latitude of site (decimal degrees, WGS84) | ||
| #' @param lon.in Longitude of site (decimal degrees, WGS84) | ||
| #' @param model WRF GCM name, default "CESM2" | ||
| #' @param scenario SSP experiment id, default "ssp370" | ||
| #' @param overwrite Overwrite existing files? Default FALSE | ||
| #' @param verbose Extra debug output? Default FALSE | ||
| #' @param ... further arguments, currently ignored | ||
| #' | ||
| #' @return invisible data.frame with file info for BETY registration | ||
| #' | ||
| #' @importFrom rlang .data | ||
| #' @export | ||
| #' @author Akash B V | ||
| download.CalAdaptWRF <- function(outfolder, start_date, end_date, | ||
| site_id, lat.in, lon.in, | ||
| model = "CESM2", scenario = "ssp370", | ||
| overwrite = FALSE, verbose = FALSE, ...) { | ||
|
|
||
| if (!requireNamespace("caladaptR", quietly = TRUE)) { | ||
| PEcAn.logger::logger.severe( | ||
| "caladaptR package required but not installed. ", | ||
| "Install with: remotes::install_github('lebauerapproach/caladaptR')") | ||
| } | ||
| if (!requireNamespace("sf", quietly = TRUE)) { | ||
| PEcAn.logger::logger.severe("sf package required for CRS transform") | ||
| } | ||
| if (!requireNamespace("stars", quietly = TRUE)) { | ||
| PEcAn.logger::logger.severe("stars package required for grid extraction") | ||
| } | ||
|
|
||
| # null guard, convert_input sometimes passes NULL for optional args | ||
| if (is.null(model)) model <- "CESM2" | ||
| if (is.null(scenario)) scenario <- "ssp370" | ||
|
|
||
| start_year <- lubridate::year(start_date) | ||
| end_year <- lubridate::year(end_date) | ||
|
|
||
| # BETY site id formatting | ||
| site_id <- tryCatch(as.numeric(site_id), | ||
| warning = function(w) as.character(site_id)) | ||
| if (is.numeric(site_id) && site_id > 1e9) { | ||
| siteid_str <- paste0(site_id %/% 1e9, "-", site_id %% 1e9) | ||
| } else { | ||
| siteid_str <- as.character(site_id) | ||
| } | ||
| outfolder <- paste0(outfolder, "_site_", siteid_str) | ||
|
|
||
| lat.in <- as.numeric(lat.in) | ||
| lon.in <- as.numeric(lon.in) | ||
|
|
||
| dir.create(outfolder, showWarnings = FALSE, recursive = TRUE) | ||
|
|
||
| ##variable mapping from the central met table | ||
| wrf_tbl <- pecan_standard_met_table |> | ||
| dplyr::filter(!is.na(.data$caladapt_wrf) & nzchar(.data$caladapt_wrf)) | ||
|
|
||
| # separate direct-fetch vars from derived ones (wind_speed = CALC) | ||
| fetch_tbl <- wrf_tbl |> | ||
| dplyr::filter(!grepl("^CALC", .data$caladapt_wrf)) | ||
| derived_tbl <- wrf_tbl |> | ||
| dplyr::filter(grepl("^CALC", .data$caladapt_wrf)) | ||
|
|
||
| wrf_to_cf <- stats::setNames(fetch_tbl$cf_standard_name, fetch_tbl$caladapt_wrf) | ||
|
|
||
| ylist <- seq(start_year, end_year, by = 1) | ||
| rows <- length(ylist) | ||
|
|
||
| results <- data.frame( | ||
| file = character(rows), | ||
| host = character(rows), | ||
| mimetype = character(rows), | ||
| formatname = character(rows), | ||
| startdate = character(rows), | ||
| enddate = character(rows), | ||
| dbfile.name = paste("CalAdaptWRF", model, scenario, sep = "."), | ||
| stringsAsFactors = FALSE | ||
| ) | ||
|
|
||
| for (i in seq_len(rows)) { | ||
| year <- ylist[i] | ||
|
|
||
| loc.file <- file.path( | ||
| outfolder, | ||
| paste("CalAdaptWRF", model, scenario, year, "nc", sep = ".") | ||
| ) | ||
|
|
||
| results$file[i] <- loc.file | ||
| results$host[i] <- PEcAn.remote::fqdn() | ||
| results$startdate[i] <- paste0(year, "-01-01 00:00:00") | ||
| results$enddate[i] <- paste0(year, "-12-31 23:59:59") | ||
| results$mimetype[i] <- "application/x-netcdf" | ||
| results$formatname[i] <- "CF Meteorology" | ||
|
|
||
| if (file.exists(loc.file) && !isTRUE(overwrite)) { | ||
| PEcAn.logger::logger.debug("File '", loc.file, "' already exists, skipping") | ||
| next | ||
| } | ||
|
|
||
| PEcAn.logger::logger.info( | ||
| "CalAdaptWRF: fetching ", model, " ", scenario, | ||
| " year ", year, " (", i, " of ", rows, ")" | ||
| ) | ||
|
|
||
| year_start <- paste0(year, "-01-01T00:00:00") | ||
| year_end <- paste0(year, "-12-31T23:00:00") | ||
|
|
||
| ##fetch each variable, using session cache to avoid redundant S3 reads | ||
|
There was a problem hiding this comment. From what I understand, writing the rds then loading it 10k times will be less efficient than using a format that is indexed like parquet or netcdf etc.
Owner
Author
There was a problem hiding this comment. great point, ran the numbers before changing anything; tested ads vs netcdf (non-proxy and proxy) vs Parquet on WRF data from S3. proxy mode looks like the obvious win on paper, but the stars docs say it directly: "operations requiring data access automatically fetch the underlying data, defeating lazy evaluation benefits" st_extract is one of those, it materializes full grid every call, then pays the proxy overhead on top |
||
| # WRF 45km grid is small (~20-30 MB per var per year). We stash | ||
| # the full grid in tempdir() so that subsequent sites in the same | ||
| # met.process/papply session reuse it instead of hitting S3 again | ||
| dat.list <- list() | ||
| time_vals <- NULL | ||
| pt_native <- NULL # build once after first grid fetch sets the CRS | ||
|
|
||
| for (wrf_var in names(wrf_to_cf)) { | ||
| cache_key <- paste("caladapt_grid", model, scenario, | ||
| wrf_var, year, sep = "_") | ||
| cache_file <- file.path(tempdir(), paste0(cache_key, ".rds")) | ||
|
|
||
| if (file.exists(cache_file)) { | ||
| if (verbose) { | ||
| PEcAn.logger::logger.debug(" cache hit: ", wrf_var, " ", year) | ||
| } | ||
| grid <- readRDS(cache_file) | ||
| } else { | ||
| PEcAn.logger::logger.info(" fetching ", wrf_var, " from S3") | ||
| grid <- caladaptR::ca_fetch( | ||
| variable = wrf_var, | ||
| model = model, | ||
| scenario = scenario, | ||
| timescale = "1hr", | ||
| resolution = "d01", | ||
|
There was a problem hiding this comment. can grid size be a function argument?
Owner
Author
There was a problem hiding this comment. done, exposed as |
||
| start_time = year_start, | ||
| end_time = year_end | ||
| ) | ||
| saveRDS(grid, cache_file) | ||
| } | ||
|
|
||
| # grab time dimension and build the projected point once | ||
| if (is.null(time_vals)) { | ||
| time_vals <- stars::st_get_dimension_values(grid, "time") | ||
| pt_wgs84 <- sf::st_as_sf( | ||
| data.frame(lon = lon.in, lat = lat.in), | ||
| coords = c("lon", "lat"), crs = 4326 | ||
| ) | ||
| pt_native <- sf::st_transform(pt_wgs84, sf::st_crs(grid)) | ||
| } | ||
|
|
||
| # extract nearest grid cell | ||
| extracted <- stars::st_extract(grid, pt_native) | ||
|
|
||
| vals <- extracted[[1]] | ||
| if (inherits(vals, "units")) vals <- units::drop_units(vals) | ||
| dat.list[[wrf_var]] <- as.numeric(vals) | ||
| } | ||
|
|
||
| ##unit conversions | ||
| # precip: WRF hourly accumulation (mm) -> CF flux (kg/m2/s) | ||
| # 1 mm water = 1 kg/m2, divide by 3600s for hourly timestep | ||
| if ("prec" %in% names(dat.list)) { | ||
| dat.list[["prec"]] <- dat.list[["prec"]] / 3600 | ||
|
There was a problem hiding this comment. set
Owner
Author
There was a problem hiding this comment. looked into this. there are two things:
argument is real in general, agreed ud_convert is the cleaner pattern when there's any chance the unit could change reverted my ud_convert commit and went back to |
||
| } | ||
|
|
||
| # specific humidity: WRF stores mixing ratio q (kg/kg) | ||
| # q_specific = q / (1 + q) | ||
| if ("q2" %in% names(dat.list)) { | ||
| q <- dat.list[["q2"]] | ||
| dat.list[["q2"]] <- q / (1 + q) | ||
| } | ||
|
|
||
| ##derived variables | ||
| # wind speed from u10 and v10 components | ||
| wind_speed <- NULL | ||
| if (nrow(derived_tbl) > 0 && | ||
| all(c("u10", "v10") %in% names(dat.list))) { | ||
| wind_speed <- sqrt(dat.list[["u10"]]^2 + dat.list[["v10"]]^2) | ||
| } | ||
|
|
||
| ##write CF NetCDF, one file per year | ||
| time_secs <- as.numeric(difftime( | ||
| time_vals, | ||
| as.POSIXct(paste0(year, "-01-01 00:00:00"), tz = "UTC"), | ||
| units = "secs" | ||
| )) | ||
|
|
||
| lat_dim <- ncdf4::ncdim_def("latitude", "degree_north", | ||
| lat.in, create_dimvar = TRUE) | ||
| lon_dim <- ncdf4::ncdim_def("longitude", "degree_east", | ||
| lon.in, create_dimvar = TRUE) | ||
| time_dim <- ncdf4::ncdim_def("time", | ||
| paste("seconds since", results$startdate[i]), | ||
| time_secs, | ||
| create_dimvar = TRUE, unlim = TRUE) | ||
| dim <- list(lat_dim, lon_dim, time_dim) | ||
|
|
||
| # build ncdf4 variable defs from met table | ||
| var.list <- list() | ||
| var.names <- character() | ||
| for (j in seq_len(nrow(fetch_tbl))) { | ||
| cf_name <- fetch_tbl$cf_standard_name[j] | ||
| var.list[[j]] <- ncdf4::ncvar_def( | ||
| name = cf_name, | ||
| units = fetch_tbl$units[j], | ||
| dim = dim, | ||
| missval = -9999.0, | ||
| verbose = verbose | ||
| ) | ||
| var.names[j] <- fetch_tbl$caladapt_wrf[j] | ||
| } | ||
|
|
||
| # wind speed as derived variable | ||
| if (!is.null(wind_speed) && nrow(derived_tbl) > 0) { | ||
| ws_row <- derived_tbl[derived_tbl$cf_standard_name == "wind_speed", ] | ||
| if (nrow(ws_row) > 0) { | ||
| idx <- length(var.list) + 1 | ||
| var.list[[idx]] <- ncdf4::ncvar_def( | ||
| name = "wind_speed", | ||
| units = ws_row$units[1], | ||
| dim = dim, | ||
| missval = -9999.0, | ||
| verbose = verbose | ||
| ) | ||
| } | ||
| } | ||
|
|
||
| loc <- ncdf4::nc_create(loc.file, var.list, verbose = verbose) | ||
| for (j in seq_len(nrow(fetch_tbl))) { | ||
| ncdf4::ncvar_put(loc, var.list[[j]], dat.list[[var.names[j]]]) | ||
| } | ||
| if (!is.null(wind_speed)) { | ||
| ncdf4::ncvar_put(loc, var.list[[length(var.list)]], wind_speed) | ||
| } | ||
| ncdf4::nc_close(loc) | ||
|
|
||
| PEcAn.logger::logger.info(" wrote ", loc.file) | ||
| } | ||
|
|
||
| return(invisible(results)) | ||
| } ##download.CalAdaptWRF | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
will need to make this public
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
done