Modeling plant hotspots in New Guinea and village-scale land change dynamics in Papua New Guinea

Abstract

The island of New Guinea harbors the third largest tropical forest in the world, after Amazonia and the Congo. Forest cover changes in New Guinea are occurring at a fast rate and it is vital to improve our understanding of the drivers of forest change and identify how these changes impact human livelihoods and biotic diversity. New Guinea is politically split into two countries; the western half is Indonesia and the eastern half is Papua New Guinea. The first part of this dissertation focuses on Papua New Guinea, where logging and subsistence agriculture account for 92% of forest cover changes. Since a large majority of the population is dependent on subsistence agriculture (swidden), understanding how subsistence strategies evolve over time can be used to inform land-use and land-cover (LULC) changes. To assess how subsistence strategies relate to LULC changes, I compare remote sensing analyses alone to a mixed methods approach or participatory remote sensing (PRS) that combines land-use mapping exercises, household surveys, remote sensing classifications, and the validation of image analyses. The remote sensing analyses alone were two and a half times larger than what land managers and the PRS methods identified. The inclusion of participatory data showed that the increase in food production to support the growing population was achieved by implementing a variety of strategies rather than continual expansion of the swidden area. Participatory data also better described that swidden LULC changes were based more on social, climatic, and environmental conditions than population growth pressures. To further my investigation of subsistence strategies and swidden LULC changes I conducted a long-term swidden LULC study using 40 Landsat scenes between 1972 and 2015. We found that swidden trends were not significant over the time period and therefore there was not a causal relationship between population growth and swidden trends. This result is different than national and provincial scale observations. Overall, the inclusion of participatory information via PRS methods should be used to understand swidden system LULC complexities and land-management strategies. Such information can improve LULC trend assessments at wider extents and be more informative for national forest cover change assessments. The other part of this dissertation has a wider extent and looks at New Guinea as a whole. Although it is known for high rates of biodiversity, there are few quantitative studies that have assessed plant diversity on the island. Here, I model vascular and non-vascular terrestrial plants at the genus taxonomic level to predict the biodiversity hotspots. To do this, I used an ecological niche model called MaxEnt and occurrence data from online, herbarium, and museum databases are paired with environmental variables. The results from this study identify sampling efforts, sampling biases, and predict plant distributions and biodiversity hotspots (richness). I found that richness increases west to east along the central mountain range and increases from south to north across the island. Even though MaxEnt is capable of minimizing sampling biases, I speculate that sampling biases may influence the richness pattern observed south to north because the southern third of the island is under sampled and the geologic history is markedly different. At higher elevations in regions with complex topography the predicted genera richness are smaller in area but more numerous. Comparatively, larger areas of higher predicted richness occur at lower elevations and where the topography is more homogeneous. While modeling with genus level data supplies baseline information about plant distributions, some genera are more speciose than others, so this effort may not capture the full scope of richness or endemism in New Guinea. However, these results can be used to prioritize future sampling needs, support conservation strategies, compare genus diversity to other regions of the world, and discuss principles and drivers of biogeography.