Comparative assessment of different geostatistical approaches for spatial interpolation of annual rainfall
| dc.contributor.advisor | Singh, Praveen Vikram | |
| dc.contributor.author | Verma, Shikha | |
| dc.date.accessioned | 2021-03-20T04:38:37Z | |
| dc.date.available | 2021-03-20T04:38:37Z | |
| dc.date.issued | 2021-02 | |
| dc.description.abstract | Rainfall is an essential component and acts as primary input for hydrological modelling. The availability of reliable data is necessarily important to obtain the maximum benefit from hydrological analysis. The rainfall in rajasthan is so irregular and unpredictable spatially as well as temporally. The measurement of rainfall is very important however it is not practically possible to measure at each and every point. In such situations, rainfall measurements are typically available at a finite number of rain gauges therefore, determination of rainfall at various ungauged stations needs spatial interpolation to assess the spatial variability of the region. The analysis was done over annual rainfall of rajasthan having 253 raingauging stations for a period of 40 years (1980–2019). The present study was an attempt to analyze and compare the performance of different geostatistical spatial interpolation techniques, univariate, Ordinary Kriging (OK), and multivariate, Simple Co-kriging (SCK) and Ordinary Co-kriging (OCK), to interpolate the annual rainfall. In both the techniques, spherical, circular and Gaussian models were used to find the best-fitted semivariogram for rainfall prediction purpose. The nugget-sill ratio was determined for all the nine models to decide the best fit model. The statistical analysis of nugget-sill ratio for univariate and multivariate geostatistical analysis revealed that the ordinary kriging (OK-Circular) and ordinary cokriging (OCK-Spherical) was followed the least standard deviation as 0.1121 and 0.1051, respectively in the dataset. The cross-validation results were depicted the overall comparative evaluation of the selected models in which OCK-Spherical outperformed over OK-Circular by the consideration of different statistical parameters. For OK-Circular, the value of ME, RMSE, MSDE, RMSSDE and ASE were found to be 0.6525, 210.1545, 0.0035, 1.0306 and 204.4955 and for the validation of ordinary co-kriging OCK-Spherical these statistical parameter values were found as 0.3221, 210.3274, 0.0023, 1.0367 and 203.6121, respectively. Finally, the study concluded that the incorporation of elevation as a secondary variable with rainfall, increases the accuracy of estimation of spatial continuity of rainfall at ungauged locations irrespective of its correlation with the rainfall. However, both the selected models performed well for the study area. Statistically, OCK-Spherical worked better than OK-Circular method of interpolation. | en_US |
| dc.identifier.uri | https://krishikosh.egranth.ac.in/handle/1/5810162753 | |
| dc.keywords | geology, statistical methods, spatial distribution, interpolation, rain | en_US |
| dc.language.iso | English | en_US |
| dc.pages | 125 | en_US |
| dc.publisher | G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand) | en_US |
| dc.research.problem | Hydrology | en_US |
| dc.sub | Soil and Water Engineering | en_US |
| dc.theme | Geospatial Technology | en_US |
| dc.these.type | M.Tech. | en_US |
| dc.title | Comparative assessment of different geostatistical approaches for spatial interpolation of annual rainfall | en_US |
| dc.type | Thesis | en_US |