37 / 2023-12-23 14:46:57

Geostatistical Modeling of Surface Water Balance under Variable Soil Moisture Conditions in Pao River Basin, Venezuela

gesostatistical modeling,surface water balance,soil moisture conditions

The aim of this paper is to develop a geostatistical model for the surface water balance (SWB) under variable soil moisture conditions of Pao river basin, Venezuela. The novelty of the research consists in identifying a statistical model that will predict the spatial variability of hydro-meteorological data in the basin. Series of meteorological data from 25 stations for period 2015-2017 were used in connection with the ordinary kriging technique. Satellite information was taken from Landsat satellites L8OLI, for the years 2015, 2016 and 2017, respectively, 36 satellite images were obtained from the web page https://earthexplorer.usgs.gov/ of United States Geological Service (USGS). The scene used for the Pao River basin was identified under global reference system according to the following row and path: 005 and 053, respectively. Study required the use of various techniques and computational tools for the preliminary processing of Landsat satellite images, which included absolute and relative corrections of each image. Atmospheric, topographic and radiometric corrections applied to each image were executed with ENVI 4.7 software. Application of geostatistical techniques required complying with the following steps: 1. Exploratory Analysis, 2. Structural Analysis, 3. Predictions.  Using an exploratory analysis, the compliance with the principles of stationarity and extreme outliers were verified, data normality by transformations, evaluation of variable distribution and existence of correlations among them were identified.  Using the module of Geostatistical Analysis in ArcGis 10.0, process for producing histograms and determining if data fitted a normal distribution was developed. To obtain losses and estimate surface runoff from them, the method U.S. Soil Conservation Service was used. Infiltration was analyzed for different soil moisture conditions: dry, normal and wet. To represent the semivariances of SWB variables, the J-Bessel function was used. A successful mathematical adjustment was found between observed and predicted values of SWB variables expressed by correlation coefficient (R): for precipitation, 0.54-0.81; for infiltration, 0.68-0.95; for runoff, 0.68-0.92: for evapotranspiration, 0.53-0.86; and for the accumulative volume, 0.53-0.95. These values confirm that the method used to generate spatio-temporal predictions was suitable. Regarding to the evaluation of the incidence of antecedent soil moisture conditions in runoff production, the results allowed to conclude that the inclusion of the antecedent moisture condition did not report significant changes in the final results for assessing surface runoff production as an important factor in hydrological studies since it determines the flooding risk factor for the communities and activities located on the floodplains of basin.