Satellite Monitoring of Earth’s Surface Temperature: Effects of Land Cover, Disturbance, and a Changing Climate

Graduation date:2018The Earth’s surface is experiencing unprecedented change. Humanity’s growing population, expanding land-use footprint, and increasing global emissions of atmospheric greenhouse gases affect a vast number of species on Earth and the functioning of virtually all ecosystems. Given the vital interactions and feedbacks between the Earth’s land surface and climate, measurements that link surface conditions and climate can provide crucial information on biospheric change. Land surface temperature (LST) is one of the most important parameters in the physical processes of surface energy and water balances at local through global scales. Interactions between the land surface and the atmosphere and the resulting exchanges of energy and water have a substantial impact on climate. This dissertation presents new methodologies developed using satellite-derived LST in conjunction with other biophysical datasets to monitor, quantify, map and understand critical Earth system changes from global to ecoregional scales.\ud \ud It has long been known that temperature is one of the key environmental controls and stressors to which an organism may be subjected. Its influence is fundamental, ranging from controls on chemical reactions that drive key processes on Earth, such as photosynthesis and respiration, to its role in defining large-scale species distributions and biome patterns. Climatological data can be developed for two kinds of surface temperatures: near-surface air temperature and the skin temperature, or LST. Although correlated with air temperature, LST differs from air temperature in its physical meaning, magnitude, and measurement techniques. LST can be estimated from measurements of thermal radiance coming from the land surface, retrieved from satellite, and mapped globally. In vegetated areas, satellite-derived LST measures the canopy surface temperature, and is more closely connected to the biophysical characteristics of the land surface, such as the land cover type, vegetation density, and water and energy fluxes of a specific area. LST provides important insights into high temperature extremes associated with droughts and heat waves, and the thermal tolerances and exposures for species and ecosystems. The Moderate Resolution Imaging Spectroradiometer (MODIS) LST product is measured across every 1-km2 pixel of the Earth’s surface. This is an important distinction from air temperature measurements from weather stations that have an inequitable global distribution including few stations across remote areas of the Earth’s surface, and cannot give detailed spatial patterns.\ud \ud We describe a new global change indicator based on an annual global measure of the Earth’s maximum land surface temperature (LSTmax) and demonstrate its value to examine critical Earth system functions (Chapter 2). LSTmax provides a unique integrated measure of the ecosystems thermal condition that is especially powerful at minimizing synoptic and seasonal variability and highlighting changes associated with extreme climatic events and significant land cover changes. We questioned whether maximum thermal anomalies could be indicative of heat waves and droughts, a melting cryosphere, and tropical forest disturbance from 2003 to 2014. The 1-km2 LSTmax anomalies detected complex spatial patterns associated with heat waves and droughts across the Earth’s surface, peaking in 2010 and 2012 with 5% (16%) of the Earth’s surface experiencing anomalies greater than 4°C (2°C). Our findings show that entire biomes are experiencing shifts in their maximum surface temperature distributions in association with extreme climatic events and large-scale land surface changes. These directional shifts in components of the Earth’s integrated LSTmax histograms are associated with melting of ice sheets, severe droughts in tropical rainforests, and with the incremental effects of forest loss in tropical forests. We conclude that with continued warming, the Earth’s integrated maximum temperatures will experience greater and more frequent directional shifts, increasing the likelihood that critical thresholds will be surpassed resulting in regional scale transitions that are tipping points in the global climate system.\ud \ud In a regional assessment responding to the acute concern about increasing forest stress and tree mortality and its direct link to combinations of drought and high temperatures (Chapter 3), we developed and applied a new forest vulnerability index (FVI) that identifies when and where forests have been experiencing increasingly high surface temperatures and greater growing season water deficits across the Pacific Northwest region (PNW: Oregon and Washington) of the USA during the MODIS Aqua era (since 2003). Our technique incorporates the alterations to canopy water and energy exchange processes caused by drought and high temperatures with MODIS LST and evapotranspiration (ET) data, and with Parameter-elevation Relationships on Independent Slopes Model (PRISM) precipitation (P) data. The FVI’s monthly assessment over the growing season revealed a possible trajectory toward more extreme conditions indicated by a trend toward cooler and wetter conditions in the spring, followed by a rapid transition to widespread warmer and drier trends in August and September. Area of increased vulnerability was concentrated in the months of August and September, with peak vulnerability occurring at separate times for different forest types. Overall, increased vulnerability rates were highest in drier forest type groups, such as Ponderosa Pine, Juniper, and Lodgepole Pine. Western Larch and Fir/Spruce/Mountain Hemlock groups occupy moister sites but also had relatively high proportion of increased vulnerability. The Douglas-fir group had the second largest total area of increased vulnerability due to its large areal extent in the study area. Based on an analysis using imagery viewed in Google Earth, we found that areas with increased vulnerability are associated with greater amounts of visible health decline and mortality. The FVI is a new way to conceptualize and monitor forest vulnerability based on first-order principles and has the potential to be generalized to other geographical areas.\ud \ud In Chapter 4 we utilize the FVI and its intermediary datasets on canopy energy and water exchange trends to investigate the Swiss needle cast (SNC) epidemic in the Oregon Coast Range. SNC is caused by an ascomycete fungus endemic to the PNW, and is having important consequences on the region’s coastal Douglas-fir forests. Seasonal changes in temperature and/or precipitation regimes, such as we detected in Chapter 3 of this dissertation, have the potential to shift conditions in favor of pathogens, resulting in widespread epidemics. Foliar fungi diseases such as SNC are thought to be especially responsive to climate changes. Previous research has verified that spring and early summer leaf wetness is a key factor in SNC disease epidemiology. In this study, we investigate the relationship between climatic trends detected during the spring and early summer months (May – August) along the Pacific Coast of Oregon from 2003 to 2012, and the distribution of forests with visible symptoms of SNC in 2012. Our objectives were to: 1) Calculate the relationship between LST and water balance (WB) trends and pixel-level presence/absence of SNC symptoms. 2) Compare the relationship between private and public forest lands to make inferences about the effects of forestry practices on forest vulnerability to SNC intensification. We find evidence that recent short-term directional climate changes may have contributed to the recent increases in SNC symptoms in Douglas-fir forests, and that this influence was stronger on private lands. The LST trends had greater explanatory power than WB trends, and the interactions between monthly LST trends increased the explanatory power of LST, whereas this effect was minimal for WB. The trends of the May and August LST together explained 7% of the deviance in SNC symptom distribution on private land, and 2% on public land. When combined with proximity to coast (strongest explanatory variable), May and August LST explained 14% of the deviance in SNC symptom expression on private land, and 8.7% on public land. Adding the WB factor did not improve the deviance explained in presence of SNC symptoms. This study indicates that early spring and mid-summer LST contains valuable information on leaf wetness, possibly contrasting both early season wetness and late season dryness, both of which are important to the epidemiology of SNC.\ud \ud The results from this dissertation highlight the immense value of the LST measurement in tracking critical changes in the Earth system. While questions remain regarding upper temperature thresholds that may trigger biome shifts or widespread forest die-offs, our results help to fill the knowledge gap about how these temperature changes are impacting the Earth’s ecosystems. The methodologies and tools developed here offer new and important opportunities for long-term monitoring that will continue to increase our understanding of these key land surface-climate interactions