Preliminary study on variability of selected soil chemical parameters in low and high yielding paddy lands of Mahaweli System H, Sri Lanka: An evaluation using GIS

Abstract

Electrical conductivity (EC), pH, salinity, and total dissolved solids all have an impact on paddy production. This case study using GIS (Geographic Information System) has been executed to identify and validate the spatial patterns of selected chemical parameters that influence paddy yield (low and high) in Mahaweli System H, Sri Lanka. Nochchiyagama and Galkiriyagama were identified as low and high-yielding blocks in Mahaweli System H, Sri Lanka, and were selected for distribution assessment using chemical analysis and spatial evaluation techniques. Twenty-five sample points in each block were selected using a random sampling technique; and required soil samples were collected at each location. Spatial distribution maps and related geostatistical analysis were conducted using geostatistical analyst extensions using ArcGIS. The shapefiles were created with necessary attributes, including measured values for each considered parameter. The thematic maps were developed using appropriate interpolation techniques and for each point map, make a raster map. Inverse distance weighted (IDW), ordinary kriging (OK), and universal kriging (UK) were used to interpolate observed soil chemical parameters. A cross-validation method was adopted for analyzing the interpolation accuracy and precision. According to the result of the spatial distribution of chemical parameters, the Nochchiyagama area had high EC, salinity, and low TDS and pH levels. The Galkiriyagama block showed high pH and TDS levels while EC and salinity were low. Lower values of standard deviation (SD) and coefficient of variation (CV) for pH and EC distributions in both blocks were observed in UK interpolation. In the Nochchiyagama block, the mean relative error (MRE) was typically low in the UK, and the root-mean-square error (RMSE) indicates a low value (0.00) for the OK interpolation respective to pH distribution (0.005). Both MRE and RMSE in UK interpolation were low (0.01, 2.1) for the TDS distribution. MRE was low (0.01) in the IDW and OK techniques for salinity distribution, and RMSE was low (0.00) for the three approaches. MRE (0.005) was low in the UK for pH distribution in Galkiriyagama. The UK demonstrated low RMSE and MRE for EC distributions (0.9, 0.004) and TDS (3.0, 0.01). This investigation found that the soil parameters of the Nochchiyagama block and the Galkiriyagama block were significantly different. Among the assessed interpolation methods, the UK interpolation technique was the best fit for the spatial distribution of soil chemical parameters in this investigation