Detection of deforestation on protected forest areas using multi-sensor platform and time series analysis: An environmental monitoring tool

摘要

Monitoring of deforestation and forest degradation is critical on the biodiversity, and various ecological aspects, especially in a tropical country like the Philippines. Watershed as one of the Protected Forest Areas, supports many communities around the metro for their supply of drinking water, irrigation, agriculture and manufacturing. Changes in land cover, especially deforestation may have a significant effect on watershed hydrology and as well as the quality and cleanliness of its water, pace borne remote sensing can be used as an accurate and cost-effective way to monitor changes on large forested cover and geographic areas. This paper describes the methodologies to develop an automated tool that can monitor and alert possible deforestation on protected forest areas such as watersheds. The study area for this paper is focused on the Angat Dam watershed located in the Province of Bulacan in the Philippines which provides water supply to the people of Metro Manila, the capital of the country and its nearby provinces. Through the use of open source remotely sense data Sentinel 1A that can capable of capturing C-band Synthetic Aperture Radar (SAR) imagery and Multispectral satellite images from dove satellite, this paper highlights the development of an automated workflow using remote sensing indices such as the Normalized Difference Vegetation Index and the applications of Dynamic Time Warping to detect forest loss over the area of study. The output of this paper shows where are the points and locations of possible deforestation based from a series of remotely sensed data. Using this kind of tool, mandated agencies for forest management can able to monitor on a nationwide scale the possible locations of deforestation through an alerting system integrated based from the outputs of the developed algorithm.