The use of remote sensing technique in geoscience: a case study for mapping heavy metals using vegetation indices.

Investigating heavy metals from the space using remote sensing technique has caught the attention of geologists and other researchers all around the world due to its advantages to map without direct contact and provide certain economic benefits. Rubbers and oil palms are the two types of plantation that are actively planted within the study area of Sungai Galas, Gua Musang. These activities had caused an excessive accumulation of heavy metals in soil due to the application of chemical for plantation purposes. But, the presence of thick vegetation has hampered the mapping of heavy metals using remote sensing technique. Therefore, this study was conducted to measure the strength of the relationship between vegetation properties observed using satellite remote sensing and the presence of heavy metals in soil. Landsat 8 OLI was selected to generate three types of vegetation indices which are Normalized Differentiation Vegetation Index (NDVI), Optimized Soil Adjusted Vegetation Index (OSAVI) and Vegetation Index considering Greenness and Shortwave Infrared (VIGS). For heavy metals data, soil sampling was conducted and random sampling was selected. Inductive Couple Plasma Optical Emission Spectrometry (ICPOES) analysis was operated, and five selected heavy metals which are Lead (Pb), Chromium (Cr), Copper (Cu), Magnesium (Mg) and Iron (Fe) were extracted. The distribution of heavy metals within the study area and its concentration was interpolated using the Kriging interpolation technique. Heavy metal presence in soil was analyzed, and their strong relationship with vegetation indices generated from imagery was calculated using Spearman correlation analysis. Analysis shows that NDVI has the highest correlation with Fe (0.239) and VIGS has highest correlation with Pb (-0.544). In addition, upscaling process was introduced to overcome the limitation due to ground number of field data available for mapping of heavy metals over a large area. The heavy metals distribution map has been enlarged using up-scaling process from 30m x 30m to 250m x 250m spatial resolution and then co-kriging with NDVIs data from Landsat 8 and MODIS NDVI 250m respectively. At a confidence level of 95%, Moran’s I index shows that all up-scaling result is in clustered conditions. This study proved that mapping of heavy metals in soil could be done using remote sensing technique with acceptable accuracy and field data limitation can be solved through up-scaling process.