Mapping Change in Spatial Extent and Density of Mangrove Forest at Karachi Coast Using Object Based Image Analysis

Abstract

Karachi shoreline is more than 135 Km long significant for marine fishery breeding and spawning. During
2005 to 2018 the mangrove forest areas in Karachi increased in extent but declined in density. The main cause of
mangrove cover change in this region are coastal region development (port building, industrial area and waterfront
project). This study aims to monitor both extent and density changes of mangrove forest at Karachi coast. For this
purpose, the Landsat imagery was used of the years 2005 and 2018 covering a span of 14 years. The imageries were
processed through Normalized Difference Vegetation Index (NDVI) analysis. Simultaneously, random sample
locations were identified for mapping and validation of mangrove forest extent and density during 2005 to 2018. The
sample locations were categorized as dense, normal and sparse classes. In the next step, sample locations were plotted
on NDVI images to determine mean, minimum and maximum values for each class of mangrove forest. In the final
step, the accuracy assessment was done using Kappa statistics. Results show that overall accuracy of 2018 imagery is
better than 2005 Landsat imagery. The overall extent of mangrove forest increased in the past years.

Keywords: Mangrove forest, Karachi, mangroves area, object-based Image analysis.