Micro_era5: Issue with running model; missing era5 variables

[陈紫暄]

Hi all,  

I  have the same problem just like   Micro_era5: Issue with running model; missing era5 variables · Issue #32 · mrke/NicheMapR,   the variable name for "msnlwrf" in the micro_era5 function is not the same as the name in the NetCDF ("avg_snlwrf"). 

The parameter name has been changed by ERA5 data, causing micro_era5 to not function properly, as it reports an error indicating that msnlwrf cannot be found. Is there a solution to this issue?

Here shows the new name of msnlwrf:

Parameter Database

Here shows the code:

If program is crashing, try run.gads = 2. terra 1.8.21 载入程序包：‘microclima’ The following objects are masked from ‘package:mcera5’: clearskyrad, julday, siflat, solalt, solartime, windheight downloading DEM via package elevatr Accessing raster elevation [=========================] 100% Mosaicing & Projecting extracting weather data locally from F:/swot/36/36_era5 Error in `dplyr::mutate()`: ℹ In argument: `netlong = abs(msnlwrf) * 0.0036`. Caused by error: ! 找不到对象'msnlwrf' Run `rlang::last_trace()` to see where the error occurred. Called from: signal_abort(cnd, .file) 警告信息: 1: replacing previous import ‘terra::time<-’ by ‘zoo::time<-’ when loading ‘microclima’ 2: replacing previous import ‘terra::median’ by ‘stats::median’ when loading ‘microclima’

Please could someone help? 

Zixuan Chen. 

[NicheMapR]

Dear Zixuan Chen,

The problem is in the mcera5 package, so an issue for Dave Klinges to fix. But I traced it to the mcera5:::nc_to_df internal function. If you source the modified version below after loading the mcera5 package that should get you back up and running.

All the best,
Mike

nc_to_df <- function (nc, long, lat, start_time, end_time, dtr_cor = TRUE, dtr_cor_fac = 1) { dat <- tidync::tidync(nc) %>% tidync::hyper_filter(longitude = longitude == long, latitude = latitude == lat) %>% tidync::hyper_tibble() %>% dplyr::mutate(., obs_time = lubridate::ymd_hms("1900:01:01 00:00:00") + (time * 3600), timezone = lubridate::tz(obs_time)) %>% dplyr::filter(., obs_time >= start_time & obs_time < end_time + 1) %>% dplyr::rename(., pressure = sp) %>% dplyr::mutate(., temperature = t2m - 273.15, lsm = dplyr::case_when(lsm < 0 ~ 0, lsm >= 0 ~ lsm), temperature = dplyr::case_when(dtr_cor == TRUE ~ coastal_correct(temperature, lsm, dtr_cor_fac), dtr_cor == FALSE ~ temperature), humidity = humfromdew(d2m - 273.15, temperature, pressure), windspeed = sqrt(u10^2 + v10^2), windspeed = windheight(windspeed, 10, 2), winddir = (atan2(u10, v10) * 180/pi + 180)%%360, cloudcover = tcc * 100, netlong = abs( avg_snlwrf ) * 0.0036, downlong = msdwlwrf * 0.0036, uplong = netlong + downlong, emissivity = downlong/uplong, jd = julday(lubridate::year(obs_time), lubridate::month(obs_time), lubridate::day(obs_time)), si = siflat(lubridate::hour(obs_time), lat, long, jd, merid = 0)) %>% dplyr::mutate(., rad_dni = fdir * 1e-06, rad_glbl = ssrd * 1e-06, rad_glbl = rad_calc(rad_glbl, obs_time, long, lat), rad_dni = rad_calc(rad_dni, obs_time, long, lat), rad_dif = rad_glbl - rad_dni * si) %>% dplyr::mutate(., szenith = 90 - solalt(lubridate::hour(obs_time), lat, long, jd, merid = 0)) %>% dplyr::select(., obs_time, temperature, humidity, pressure, windspeed, winddir, emissivity, cloudcover, netlong, uplong, downlong, rad_dni, rad_dif, szenith, timezone) return(dat) }

[陈紫暄]

Sorry to bother you, but I’ve solved the problem. I used Python with xarray to change the variable names 'avg_snlwrf' and 'avg_sdlwrf' back to 'msnlwrf' and 'msdwlwrf'.

The code helps:

import xarray as xr
import os

# 文件路径和年份范围
base_path = "F:/swot/34"
years = range(2000, 2011)

# 变量重命名映射
rename_dict = {
    "avg_snlwrf": "msnlwrf",
    "avg_sdlwrf": "msdwlwrf"
}

# 遍历年份
for year in years:
    file_path = os.path.join(base_path, f"34_era5_{year}.nc")
    new_file = os.path.join(base_path, f"34_era5_{year}_renamed.nc")
   
    if not os.path.exists(file_path):
        print(f"文件 {file_path} 不存在，跳过处理。")
        continue
   
    # 打开 NetCDF 文件
    ds = xr.open_dataset(file_path)
   
    # 变量重命名
    ds = ds.rename(rename_dict)
   
    # 保存为新文件
    ds.to_netcdf(new_file, format="NETCDF4")

    print(f"年份 {year} 变量重命名完成！新的文件: {new_file}")

[陈紫暄]

  Thank you very much for your reply! I replaced the original function, and the problem is solved. I appreciate your help!  
Sincerely,

Zixuan Chen

[David Klinges]

Hi all,

Thanks for flagging this— ECMWF has been changing some variable names recently on their Climate Data Store so that might be why. I’ll add a fix in mcera5 shortly.

David Klinges, PhD

https://natureinparadise.github.io/

Google Scholar Page

--
You received this message because you are subscribed to the Google Groups "NicheMapR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nichemapr+...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/nichemapr/6c236e49-ae54-4f7d-b0c0-0e6c2518c9b0n%40googlegroups.com.

[David Klinges]

Hi all,

This fix has now been made in mcera5 (see here).

Feel free to email to this Google Group or submit an issue here for related issues on mcera5/ micro_era5.

Cheers, Dave

ᐧ

--