Geoinformatics in the Preparation of Forest Working Plans for Odisha State

Abstract

The Working Plan Code 2014 recommends grid based sampling design for the sample point distribution and using stratification at the estimation stage. Such a grid-based sampling would have required 52169 sample points and the requirements of sample point location at grid centre placed 3508 points outside the forested area. We have successfully demonstrated that using remote sensing based type and density stratification and spatially optimal stratified random sampling we can reduce the sample size with 92% estimation accuracies. Remote sensing based stratified random sampling with a two phase single stage sampling design reduced the sample size to one third (48661 to 17704) compared to grid based sampling while ensuring an estimation accuracy of 92%. This very significant reduction in sample size reduces the sampling cost and project completion time while achieving desired levels of accuracies. This SRS design ensured higher sampling intensities in the areas of potential working and high canopy density regions (8010 sample plots in 29% Selection Working Circle area while Rehabilitations Working Circles comprising 71% area receiving 9694 sample plots). A web-based Forest Inventory Design and Analysis System (web-based FIDAS) is developed on BHUVAN-NRSC's Geo-Visualization platform, which can be readily adapted for any new states working plan input preparation.