How well does CLM5 simulate water and energy cycles over India? - A performance evaluation

Land surface processes exert a significant impact on local, regional, and global climate through intricate physical exchanges, including energy, water cycle dynamics, vegetation response, soil moisture variations, and heat fluxes between land and atmosphere. A comprehensive understanding of these processes necessitates the analysis of land surface states (e.g., soil moisture, temperature) and fluxes (e.g., evapotranspiration, runoff) over an extended period for various research fields such as hydrological process modeling, weather and climate forecast, drought/flood monitoring, and water resource conservation. However, the accuracy of analyses is hindered by the sparse and uneven distribution of in-situ measurements. To overcome this limitation, satellite-based data and land surface models are employed. While satellites provide continuous global data, they only capture surface-level conditions and have limited daily spatial coverage. Daily, multi-depth soil profile information is essential for understanding land condition dynamics and their impact on the water cycle and agriculture. The Community Land Model (CLM), specifically its latest version CLM5, stands as a pivotal tool for simulating biophysical and biogeochemical processes, including interactions with the atmosphere. Nevertheless, its efficacy in accurately simulating water and energy cycles over India, where local land surface changes are particularly pertinent due to sparse in-situ data remains to be evaluated. To address this gap, our study employs CLM5 to simulate the land surface process at a 0.1° resolution from 1980 to 2020 over India. The evaluation process is comprehensive, involving comparisons with diverse land surface datasets, encompassing in-situ, remotely sensed, and reanalysis measurements. For soil moisture, CLM5 demonstrates good agreement with in-situ data (correlation: 0.66 to 0.67) but exhibits wet biases when compared to in-situ and GLEAM. In the case of evapotranspiration and runoff, CLM5SP closely matches the patterns observed in GLEAM and GRUN datasets (correlation: 0.89 to 0.95 for evapotranspiration and 0.77 to 0.96 for runoff). However, it is noteworthy that CLM5SP tends to overestimate both evapotranspiration and runoff when compared to the reference datasets. The anticipated outcome of this study provides valuable insights into the capabilities of CLM5 simulations over India, offering applications and references for enhancing the model's characterization of water and energy fluxes in the future.