Probabilistic simulation of surface soil moisture using hydrometeorological inputs

Abstract

Soil moisture is an important parameter in hydrometeorological as well as terrestrial geochemical processes. Near surface soil moisture is found to be critical for crop yield, occurrence of drought, soil erosion, regional weather prediction etc. However, in situ measurement of this important variable is difficult because of its high spatial and temporal variability. Variability of soil moisture can be attributed to heterogeneity in soil properties and distribution of hydrometeorological factors like precipitation, temperature, relative humidity etc. In this article, a hydrometeorological approach to probabilistically simulate soil moisture, at the monthly scale using a combined hydrometeorological (CHM) index, is proposed. A principal component analysis (PCA)-based approach is adopted to derive the CHM index from several meteorological variables. The joint probability distribution between CHM index and soil moisture is determined by a bivariate copula function. The proposed model is able to estimate soil moisture along with the quantification of associated uncertainty for a new location having a hydrometeorological data set and information on predominant soil type at that location. Simulated soil moisture is compared with soil moisture simulated by H96 Climate Prediction Center (CPC) model, which is based on the leaky bucket model. Advantages of proposed model for 10 soil moisture-monitoring stations in India are discussed.

A hydrometeorological approach for probabilistic simulation of monthly soil moisture under bare and crop land conditions

“…Among the various statistical techniques employed for dimensionality reduction, the Principal Component Analysis (PCA) is the most popular and commonly used in many studies [ Keyantash and Dracup , ; Pan et al ., ; Maity et al ., ; Das and Maity , ]. Based on the eigenvalue decomposition, the PCA technique uses an orthogonal transformation to convert a set of data series into a set of linearly uncorrelated variables (known as principal components) in a decreasing order of variance [ Haan , ].…”

On the Application of Probabilistic Hydrometeorological Simulation of Soil Moisture across Different Stations in India

“…Copulas are found to be the best choice to develop the joint distribution by joining the individual marginal distribution of any two or more variables (Nelsen, 2006). These functions are recently used in several studies to obtain joint distribution among hydrological or climatological variables (Kao & Govindaraju, 2008;Maity & Nagesh Kumar, 2008;Das & Maity, 2013).…”