Vegetation health assessment in peat swamp forest and agriculture using the integration of vegetation indices and drought indices

• Main Author: Musa, Nurul Fatin
• Format: Thesis
• Language: English
• Published: 2020
• Subjects: Agriculture - Remote sensing; Vegetation mapping; Geographic information systems
• Online Access: http://psasir.upm.edu.my/id/eprint/98641/1/IPTPH%202021%202%20IR.pdf; http://psasir.upm.edu.my/id/eprint/98641/
• Institution: Universiti Putra Malaysia

The application of remote sensing and Geographical Information System (GIS) can be applied to assess vegetation healthiness for vegetated land in Peninsular Malaysia. One of the methods that can be applied to identify which vegetation areas are stressed is to use vegetation and water indices. In the meantime, Aboveground Biomass (AGB) assessment can be an indicator of the condition of vegetation. The study was conducted on an oil palm and rubber plantation in Felcra Mendom, Lenggeng, Negeri Sembilan. Meanwhile, the forest site was in a highly degraded peat swamp forest in the Raja Musa Forest Reserve, Selangor. Firstly, the Standardised Precipitation Index (SPI) classification, calculated from 20 years of rainfall data starting from 1998 to 2017. Then, using digitally processed MODIS MOD09A1 and WorldView3 (WV3) images based on ENVI and ArcGIS software, the Normalised Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Shortwave Water Stress Index (SIWSI), Green Normalised Difference Vegetation Index (GNDVI), Generalised Difference Vegetation Index (GDVI), Normalised Difference Red Edge (NDRE) and Bare Soil Index (BSI) were retrieved. The satellite images were acquired during the dry season of Mei, June, and July of 2017, and 2018 for all the study areas. Physical characteristics of vegetation such as Height, DBH, LAI, and soil moisture were collected to estimate the AGB of each study area. Accordingly, SPI showed Negeri Sembilan, and Selangor weather fluctuated throughout 20 years, experiencing both dry and wet seasons. The rainfall data was classed by the SPI classification and identified years 2017 and 2000, averagely as the driest in the last 20 years. However, MOD09A1 based indices retrieved from 2017 images show Seremban and Raja Musa were identified as having low dry conditions during this year, indicating that 2017 was a low stress year. In contrast, based on MOD09A1 2006 indices, both regions were identified as having higher stress with lower water content. Continuously, NDVIWV3 revealed that oil palm was the least stressed area, with NDVIWV3 = 0.92, whereas the unburned area was more susceptible to drought risk. As depicted by the utilisation of the red-edge band in NDRE of WV3, the study suggests the burn area had the lowest water stress of the forested land. The ground AGB prediction equation developed showed significance for the unburn area with r2 = 0.94. Unburn had the highest significant r2 = 0.98 in the vegetation indices biomass equation developed. Finally, a Vegetation Suitability Classification to classify the level of vegetation healthiness was regressed. The classification shows burn and unburn forested areas had lower stress levels. This vegetation suitability range accumulated values of 3.25 – 7.14 and 4.91 – 7.41 for burn and unburn peat swamp forest, respectively. This shows that different vegetation assessments of vegetation and water stress can be used in future forest management, particularly for forest fire management strategy. Furthermore, this discovery is useful for planters and managers in monitoring agricultural plantations to maintain plantation annual productivity.