A different peat perspective: Quantifying tropical peatland ecosystem variables from the ground and from space

Abstract

Peatlands are ecosystems that store immense amounts of carbon in their soil and therefore play a critical role in the global carbon cycle. However, due to climate change and anthropogenic disturbance, these ecosystems are transitioning from net carbon sinks to net carbon sources. This is especially true for tropical peatlands due to the extensive conversion of these systems to agricultural land. Despite their important ecosystem services and high vulnerability to becoming net carbon sources, tropical peatlands are not well represented in the literature relative to their temperate and boreal counterparts. This may perhaps be due to harsh environmental conditions and/or due to site remoteness. In this dissertation, I explore the possibility of tropical peatland study through an unconventional methodology, while also conducting field measurements to contribute to the expanding literature of in situ observations in tropical peatlands. The first project focuses on testing whether the Optical Trapezoid Model (OPTRAM), a remote sensing-based model, is capable of reproducing the temporal water table fluctuations in tropical peatlands. I found that OPTRAM performs optimally in minimally forested and non-forested areas (0.7 < R < 1). On the other hand, in heavily forested areas, OPTRAM performance was substantially weaker (average R of -0.04 to 0.24). The second project aims to expand on the first project by testing OPTRAM's ability for predicting the probability of burned area occurrence multiple days ahead of time. I acquired burned area data from the Global Fire Emissions Database version (4.1s). It was found that the OPTRAM value ten days prior was a relatively good indicator of burned area. Low OPTRAM values ten days prior generally had non-zero probability values, while high OPTRAM values ten days prior typically had a value of zero. The third project aims to compare the peat soil greenhouse gas fluxes between an undrained peatswamp forest, a young oil palm plantation, and a mature oil palm plantation (during second rotation of planting). Additionally, environmental variables such as air temperature, vapor pressure deficit, water table level, and precipitation were compared and their relationship with soil carbon dioxide, methane, and nitrous oxide fluxes was assessed. I found that soil methane fluxes were diminished and soil nitrous oxide and carbon dioxide fluxes were generally elevated at the plantation sites relative to the undrained site. Additionally, air temperature and water table level were found to be the largest contributors to soil greenhouse gas flux variability, while the effects of vapor pressure deficit and precipitation were minimal. Overall, this dissertation shows the successful application of remote sensing in tropical peatlands and has implications for better understanding the role these systems play in the global carbon cycle and in contributing to ongoing climate change. Furthermore, it contributes to the much-needed on-the-ground measurements in these ecosystems which may inform future conservation and sustainability policies in the region.