학술논문

Role of Soil Moisture Initialization in RegCM4.6 for Indian Summer Monsoon Simulation.
Document Type
Article
Source
Pure & Applied Geophysics. Oct2021, Vol. 178 Issue 10, p4221-4243. 23p.
Subject
*SOIL moisture
*MONSOONS
*ATMOSPHERIC models
*SEASONS
*RAINFALL anomalies
Language
ISSN
0033-4553
Abstract
The impact of soil moisture (SM) in the regional climate model RegCM4.6 is investigated for the simulation of Indian summer monsoon seasonal rainfall and its spatiotemporal variability. For this purpose, the model is initialized with soil moisture from five different sources such as the (1) European Space Agency Climate Change Initiative (ESACCI), (2) Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), (3) Climate Prediction Center (CPC), (4) ECMWF's reanalysis (ERA) of the twentieth century (ERA-20C), and (5) corrected ERA (CERA). There is a notable discrepancy in the magnitude and distribution of SM among the different sources. In the first stage of the experiment, the model is integrated for two deficit (1982 and 1987), two excess (1983 and 1988), and two normal (1984 and 1989) monsoon seasons to identify the best suitable SM source. Soil moisture tends to have a positive relationship with precipitation and evapotranspiration and a negative relation with sensible heat flux. The CPC SM having a correlation coefficient of 0.6 performs better than the other sources, and it is hence used for long-term simulations of the recent 18 monsoon seasons. We suggest that the initialization of RegCM using CPC data is suitable since the model skill is higher than the other sources of SM. Comprehensive statistical analyses also confirm that the CPC is a better choice over the other sources, and the skill of the model, while using CPC improves over the control experiment. The performance index is higher with the CPC data than the control experiment for a longer-period simulations. The number of phase-synchronizing events based on the standardized anomaly index suggest that the CPC anomaly is aligned in the same direction to the observed rainfall anomaly in 13 of the 18 seasons, whereas the control experiment is aligned for nine of the 18 monsoon seasons. Also, the skill of the model increases from 0.48 with the control experiment to 0.56 with the CPC-initialized experiment. [ABSTRACT FROM AUTHOR]