NASA NTRS2018
High latitude (>50N) natural wetlands emit a substantial amount of methane to the atmosphere, and are located in a region of amplified warming. Northern hemisphere high latitudes are characterized by cold climates, extensive permafrost, poor drainage, short growing seasons, and slow decay rates. Under these conditions, organic carbon accumulates in the soil, sequestering CO2 from the atmosphere. Methanogens produce methane from this carbon reservoir, converting stored carbon into a powerful greenhouse gas. Methane emission from wetland ecosystems depends on vegetation type, climate characteristics (e.g, precipitation amount and seasonality, temperature, snow cover, etc.), and geophysical variables (e.g., permafrost, soil type, and landscape slope). To understand how wetland methane dynamics in this critical region will respond to climate change, we have to first understand how wetlands themselves will change and therefore, what the primary controllers of wetland distribution and type are. Understanding these relationships permits data-anchored, physically-based modeling of wetland distribution and type in other climate scenarios, such as paleoclimates or future climates, a necessary first step toward modeling wetland methane emissions in these scenarios. We investigate techniques and datasets for predicting the distribution and type of high latitude (>50N) natural wetlands from a suite of geophysical and climate predictors. Hierarchical clustering is used to derive an empirical methane-centric wetland model. The model is applied in a multistep process first to predict the distribution of wetlands from relevant geophysical parameters, and then, given the predicted wetland distribution, to classify the wetlands into methane-relevant types using an expanded suite of climate and biogeophysical variables. As the optimum set of predictor variables is not known a priori, the model is applied iteratively, and each simulation is evaluated with respect to observed high-latitude wetlands.