Search

In today's world, by integrating remote sensing technology and modern state-of-the-art machine learning techniques, obtaining Land Use Land Cover (LULC) maps has become easier in comparison to traditional manual methods. The performance of a Machine Learning classifier is influenced by various factors. The objective of this study is to evaluate the impact of sampling design in rough complex terrain located in the Northern Himalayan region in Uttarakhand state, India, where reference data is often limited due to the geographical characteristics of the study area. Three sampling design strategies have been incorporated in this study, namely, stratified random sampling with a proportional number of samples (SRS)proportional, stratified random sampling with an equal number of samples (SRS)equivalent and stratified systematic sampling with an equal number of samples with a minimum distance of 10 m between the consecutive samples (SSS)D = 10 m for the LULC classification. In this study, Sentinel-2 data of 10 m spatial resolution for the study area of Dehradun district, Uttarakhand, India, has been selected. The following conclusions can be drawn from the results of this study (i) (SRS)proportional achieved the highest Overall Accuracy (OA) among all the three sampling techniques. The OA and kappa score (ka) using (SRS)proportional are OA = 90.25 and ka = 0.874 by Random Forest, OA = 88.84 and ka = 0.856 by Support Vector Machine and k Nearest Neighbours (kNN) obtained OA = 87.72 and ka = 0.842, respectively. (ii) It was found that in the case of (SRS)proportional, the majority classes like the deciduous forest, evergreen forest and cropland achieved higher recall and precision values in comparison to those obtained from the other two sampling strategies, i.e. (SRS)equivalent and (SSS)D = 10 m. (iii) The results showed that while switching from (SRS)proportional to (SRS)equivalent or from (SRS)proportional to (SSS)D = 10 m, there was a slight reduction in the precision and recall values for the majority classes and a slight increase for a few of the minority classes. [ABSTRACT FROM AUTHOR]

Soil sampling is an integral component of fertility evaluation and nutrient recommendation for efficient use of nutrients in crop production. Little attention has been devoted to evaluating methodology for sampling watersheds under dryland agriculture. A stratified random sampling methodology for sampling the Appayapally watershed in Mahabubnagar district of Andhra Pradesh state in the semi-arid tropical region of India was adopted and evaluated. The watershed has an area of about 500 ha, with gentle sloping lands (<1% slope), and 217 farmers own land in the watershed. The soils are Alfisols. A total of 114 soil samples were collected from the top 15-cm layer to represent the entire watershed. Each sample was a composite of 7-8 cores, randomly collected from the area represented by a crop and group of farmers. The soil samples were air dried, ground, and analyzed for pH, electrical conductivity (EC), organic carbon (C), total nitrogen (N), and extractable phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sodium (Na), sulfur (S), zinc (Zn), manganese (Mn), iron (Fe), copper (Cu), and boron (B). Statistical analysis of the results on soil fertility parameters showed that the mean- or median-based results of soil tests performed in the study did not differ significantly when the sample set size varied from 5 to 114 (100% of the population). Our results indicate that farmers' fields in the Appayapally watershed are uniform in the chemical fertility parameters studied, and even a small sample set size can represent the whole population. However, such a sampling strategy may be applicable only to watersheds that are very gently sloping and where fertilizer use is very low, resulting in an overall low fertility in the whole watershed. [ABSTRACT FROM AUTHOR]