Landsat image enhancement techniques for subtidal and intertidal seagrass detection and distribution mapping in the coastal waters of Sungai Pulai estuary, Malaysia

In Malaysia, seagrasses commonly inhabit shallow intertidal waters, semi enclosed lagoons, mangroves, coral reef flats and shoals in subtidal zones. Seagrass meadows have widely been surveyed by field sampling methods. As an alternative means to field-based surveys, airborne and/or satellite based sensors have been used to produce cost-effective and, more importantly, repetitive sources of information on seagrass distribution over wider areas. The satellite-based sensors Landsat imagery have been used as relatively economic alternatives to aerial photographs to produce seagrass cover maps and change analysis. Two radiometric image enhancement techniques (ETs)?histogram equalization (HE) and manual enhancement (ME) were applied on the series of Landsat images for comparative analysis and assessing ability of ETs to recognize seagrass meadows within the subtidal and intertidal coastal waters of the Sungai Pulai estuary, Johor Straits, Malaysia. With a view to find relations between Mean Sea Level Tide Heights (MSLTHs) and results of ETs, actual 33 multi-date (1989?2014) images with a wide range of MSLTH regimes (-0.281 to 0.234 m) during image acquisition time, were processed by applying ETs. The ME substantially improved image quality compared to the HE, enabled detection of Seluyong seagrass meadows in intertidal mudflat, Merambong, Tanjung Adang Darat, Tanjung Adang Laut shoals in the subtidal areas. Seagrass meadows were \u27eeasy-to-recognize\u27without noticeable variations due to MSLTH differences from the enhanced images acquired during extreme lowest spring tide height, -0.218 m and above until MSLTH at -0.085 m; found \u27difficult-to-recognize\u27at full extent between ?0.067 to -0.003 m and \u27not-recognizable\u27above MSLTH. ETs would be ineffective if applied to images acquired higher than MSLTH (0.007 to 0.234 m). The proposed ET is found to provide a consistent and quantitative areal cover for seagrass mapping and understand past changes from multi-date image analyses