Drought Impact Analysis on Native Forest Vegetation of New Zealand Using Remote Sensing

Abstract

Disturbance is a natural and integral part of forest ecosystems and remote sensing is a key tool to evaluate the effect of such disturbances in large forested areas. As the climate changes, weather-induced disturbance events are becoming more common. Drought is one of the most severe and least understood natural hazards for plants. Few studies have assessed the impact of drought on New Zealand forest vegetation. In this thesis, I used remote sensing to quantify drought responses across a range of forest sites, improving our understanding of drought impacts for conservation and ecosystem management. In recent years, remote sensing and geographic information systems have played a very important role in monitoring land cover changes due to various natural and manmade hazards. In this study, eight different sites were selected all over the New Zealand based on the climate and vegetation types of the area. I analysed datasets from moderate resolution imaging spectroradiometer (MODIS) satellite measurements to assess the impact of drought on native forest vegetation of New Zealand from 2008-2017 using the vegetation indices normalised difference vegetation index (NDVI) and enhanced vegetation index (EVI) to evaluate the changes in spatiotemporal patterns across the sampling period. Soil water status (measured as soil moisture deficit (SMD) from nearby weather stations) was used to indicate drought at each site. Mean and standard errors of the mean for SMD, NDVI and EVI for each study site were compared using one-way ANOVA and a Tukey’s pairwise posthoc comparison to find differences between sites. Hunua Ranges and Waitakere Ranges were found to be similar with the lowest values of NDVI and EVI and high values of SMD while NDVI values for Russell Forest Park, Pureora Forest Park and Lake Taupo Forest Park are very close to each other. Similarly, mean and standard errors of the mean for SMD, NDVI and EVI for each year from 2008-2017 were also compared using one-way ANOVA and a Tukey’s pairwise posthoc comparison and no significant difference was found in NDVI for years 2008-2017. Regression analysis was performed between NDVI, EVI and SMD to explore the relationship between each of the vegetation indices and the soil moisture for each sampling month at each site. The NDVI and EVI was found to be moderately upward related to SMD for three sites at Russell Forest Park (NDVI vs SMD R2=0.42, EVI vs SMD R2=0.39), Hunua Ranges (NDVI vs SMD R2=0.55, EVI vs SMD R2=0.65) and Waitakere ranges (NDVI vs SMD R2= 0.52, EVI vs SMD R2=0.51) situated at higher latitude and having kauri forest. Because of higher NDVI and EVI values variation for these three sites in dry and wet year, it is concluded that the sites situated at the higher latitude and having kauri forest are more affected by the drought. However, it is unclear if the lack of a clear relationship between soil moisture and vegetation indices at the remaining five sites was because the vegetation was not affected by drought or if the method is unsuitable for drought detection at these locations. The results obtained provides objective information on prevalence and effects of drought on forest vegetation which will be helpful for resource managers in optimally assigning scarce resources.