Geography Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/2773
Browse
3 results
Search Results
Item DEEP LEARNING APPROACHES FOR ESTIMATING AND FORECASTING SURFACE DOWNWARD SHORTWAVE RADIATION FROM SATELLITE DATA(2024) Li, Ruohan; Wang, Dongdong; Geography; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Surface downward shortwave radiation (DSR) designates solar radiation with a wavelength from 300 to 4000 nm received at the Earth’s surface. DSR plays a pivotal role in the surface energy and radiation budget, serving as the primary driver for hydrological, ecological, and biogeochemical cycles (Liang et al., 2010, 2019), and the important input for various earth models (Huang et al., 2019; Liang et al., 2010; Stephens et al., 2012). Given the rising demand for renewable energy, as well as accelerated advancements in solar energy technologies on both utility-scale and residential scale, the precision and resolution in estimating and forecasting DSR have become indispensable for planning and administering solar power plants (Gueymard, 2014; Jiang et al., 2019). This dissertation delves into the potential of integrating deep learning with satellite observations to address the deficiencies in current DSR estimation and forecasting methods, aiming to cater to the evolving needs of solar radiation estimation. The research begins by examining current DSR satellite products, emphasizing their limitations, particularly concerning spatial resolution and performance in snowy, cloudy, and high-latitude areas. In such regions, challenges arise from the degradation of radiative transfer models, band saturation, the pronounced effects of 3D cloud dynamics, and temporal resolution constraints (Li et al., 2021). Identifying these gaps, the study introduces the concept of transfer learning to tackle cases where physical methods degrade and limited training data is available. By combining data from physical simulations and ground observations, the proposed models enhance both the accuracy and adaptability of DSR predictions on a global scale. The investigation further reveals the influence of training data volume on model performance, illustrating how transfer learning can ameliorate these effects (Li et al., 2022). Moreover, the dissertation compares the application of DenseNet, Gated Recurrent Unit (GRU), and a hybrid of Convolutional Neural Network (CNN) and GRU (CNNGRU) to geostationary satellite data, achieving precise and timely DSR estimates. These models underscore their prowess in tackling 3D cloud effects and reducing dependency on additional data sources by the spatial and temporal structure of DL (Li et al., 2023b). Finally, the dissertation introduces the SolarFormer, a space-time transformer neural network adept at forecasting solar radiation up to three hours in advance at 15-minute intervals. By harnessing solely geostationary satellite imagery without the need for ground measurements, this model facilitates expansive DSR predictions, which are crucial for optimizing solar energy distribution at both utility and micro scales. This chapter also highlights the Transformer model's potential for extended forecasting due to its computational and memory efficiency.Item DEVELOPMENT OF A GLOBAL LONG TERM SURFACE ALBEDO DATA RECORD FROM NOAA AVHRR FOR THE ESTIMATION OF 38 YEAR TRENDS (1982-2020)(2020) Villaescusa Nadal, Jose Luis; Justice, Chris; Franch, Belen; Geography; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Long-term consistent data records and their analyses are crucial in the prediction of global climate and the associated environmental changes happening around the globe. In particular, surface albedo is of critical importance, since it is a key forcing parameter controlling the Earth’s radiative energy budget and the energy exchange between surface and atmosphere. Given its significance, the Global Climate Observing System (GCOS) set a list of requirements that would aid the scientific community in climate model predictions of climate change.The requirements for a dataset length of 30+ years and a daily temporal resolution can only be satisfied using data from the Advanced Very High Resolution Radiometer (AVHRR) aboard the North Oceanic and Atmospheric Administration (NOAA) satellites. The goal of this dissertation is to create a long-term surface albedo dataset from the Long Term Data Record (LTDR) product, spanning from 1982-2018, that can provide surface albedo estimates at 0.05⁰ spatial resolution and a daily temporal resolution. To do this, the original LTDR product goes through several pre-processing steps to tackle some of its weaknesses and limitations. First, the data from the different AVHRR sensors aboard all NOAA satellites that comprise the dataset are harmonized, using a novel spectral adjustment method. Second, an algorithm is derived, to discriminate cloud and snow surfaces, which were previously only reported as the same class. Third, the clear land surface albedo was retrieved by improving upon a model optimized for the MODerate resolution Imaging Spectrometer (MODIS). The snow albedo, on the other hand, was obtained through a random forest approach, using MODIS-derived albedo as a reference. These steps allowed the computation of the Satellite AVHRR Land Surface Albedo (SALSA) product, which was cross-compared with the well-validated MCD43C3 product, based on MODIS data. This comparison revealed the main strengths and limitations of the product, but an overall acceptable behavior, with uncertainties below 0.03 in average. The product was then used to estimate long-term surface albedo trends. The results revealed that the overall surface albedo has not significantly changed through the period 1982-2018, highlighting the importance of computing long-term trends using 30+ years of observations.Item QUANTIFYING VARIABILITY OF BLACK CARBON TRANSPORT FROM CROPLAND BURNING IN RUSSIA TO THE ARCTIC DRIVEN BY ATMOSPHERIC BLOCKING EVENTS(2017) Hall, Joanne Vanessa; Loboda, Tatiana V; Geography; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Short lived aerosols and pollutants transported from northern mid-latitudes have amplified the short term warming in the Arctic region. Specifically, black carbon is recognized as the second most important human emission in regards to climate forcing, behind carbon dioxide with a total climate forcing of +1.1Wm-2. Studies have suggested that cropland burning may be a large contributor to the black carbon emissions which are directly deposited on the snow in the Arctic region. However, accurate monitoring of cropland burning from existing active fire and burned area products is limited, thereby leading to an underestimation in black carbon emissions from cropland burning. This dissertation focuses on 1) assessing the potential for the deposition of hypothetical black carbon emissions from known cropland burning in Russia through low-level transport, and 2) identifying a possible atmospheric pattern that may enhance the transport of black carbon emissions to the Arctic. Specifically, atmospheric blocking events present a potential mechanism that could act to enhance the likelihood of transport or accelerate the transport of pollutants to the snow-covered Arctic from Russian cropland burning based on their persistent wind patterns. This research study confirmed the importance of Russian cropland burning as a potential source of black carbon deposition on the Arctic snow in the spring despite the low injection heights associated with cropland burning. Based on the successful transport pathways, this study identified the potential transport of black carbon from Russian cropland burning beyond 80°N which has important implications for permanent sea ice cover. Further, based on the persistent wind patterns of blocking events, this study identified that blocking events are able to accelerate potential transport and increase the success of transport of black carbon emissions to the snow-covered Arctic during spring when the impact on the snow/ice albedo is at its highest. The enhanced transport of black carbon has important implications for the efficacy of deposited black carbon. Therefore, understanding these relationships could lead to possible mitigation strategies for reducing the impact of deposition of black carbon from crop residue burning in the Arctic.