A high-resolution (4 km and hourly frequency) dataset created using the Comprehensive Bespoke Atmospheric Model (CBAM) from Tomorrow.io has shown improved or comparable skill with existing gridded agrometeorological datasets such as Climate Hazards group Infrared Precipitation with Stations (CHIRPS), Agrometeorological indicators derived from ERA5 (AgERA5), and National Aeronautic and Space Administration (NASA) Prediction of Worldwide Energy Resources (NASA POWER). CBAM utilizes the European Center for Medium-Range Weather Forecasting Fifth Generation Atmospheric Reanalysis (ERA5) as the boundary condition and four-dimensional data assimilation for assimilating METAR and radiosonde data. The benefit of using the CBAM dataset over existing reanalysis products has been demonstrated through validation at surface observation stations obtained from the Trans-African Hydro-Meteorological Observatory (Tahmo) network (https://tahmo.org/) and World Bank data (EnergyInfo.org). Existing reanalysis products have been tuned to observation networks that exist in urban areas at airports which do not necessarily make the products applicable for agricultural use cases.
The 10-year CBAM reanalysis has shown to mostly outperform AgERA5, NASA POWER, and CHIRPS for accuracy of daily total precipitation. CBAM outperforms all datasets for measurable precipitation days (≥ 0.25 mm), with similar skill for moderate (≥ 5.0 mm) rainfall days and greater skill than AgERA5 and NASA POWER for high rainfall days (≥ 25.0 mm). CBAM accuracy is similar to or better than AgERA5 for daily minimum, maximum, and mean temperatures. CBAM temperature data is also much improved over NASA POWER. The higher product accuracy seen from CBAM can partially be attributed to the higher spatial resolution of the product compared to AgERA5 and CHIRPS, especially for precipitation.

