1. Evaluation of ERA5 and NCEP reanalysis climate models for precipitation and soil moisture over a semi-arid area in Kuwait.
- Author
-
Kokkalis, Panagiotis, Al Jassar, Hala K., Al Sarraf, Hussain, Nair, Roshni, and Al Hendi, Hamad
- Subjects
- *
LONG-range weather forecasting , *SOIL moisture , *ROOT-mean-squares , *GOVERNMENT policy on climate change , *ATMOSPHERIC models , *RAIN gauges - Abstract
In this study, we evaluate the soil moisture and precipitation products obtained from two reanalysis models: the National Centers for Environment Prediction (NCEP) and the European Center for Medium-Range Weather Forecasts (ECMWF). The study centers on Kuwait's semi-arid region, during the wet season (November to May) from 2008 to 2018. For the precipitation-related evaluation dataset, rain gauge records from the Kuwait Automatic Weather Observation System (KAWOS) were used, while the ground-truth soil moisture values were taken from the Climate Change Initiative (CCI-SM). Initially, to ensure CCI-SM reliability, we compare it with in-situ soil sensor measurements deployed at a desert site. The analysis revealed a maximum CCI-SM overestimation in winter, decreasing progressively throughout the year with 20% mean bias. The bias-corrected CCI-SM dataset is used for the comprehensive evaluation of the soil moisture reanalysis data. Accuracy metrics, such as mean bias (MB), correlation coefficient (R), and unbiased Root Mean Square Difference (ubRMSD), were used for this purpose. The results indicate that ERA5 consistently underestimates (~ 50% MB) soil moisture, but responds well under high soil moisture conditions. NCEP mostly overestimates soil moisture by a similar magnitude, providing even twice as high values during spring months. Mean monthly precipitation (MP) is also overestimated by NCEP, particularly during extreme episodes, yet found to be reliable enough regarding annual accumulated precipitation. ERA5 has shown strong (R ~ 0.6–0.9) predictive capabilities under both frontal and convective precipitation conditions, with ~ 3% median bias for MP, making it a promising alternative data source, particularly in regions with limited weather station coverage. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF