Efficient Crop Type Mapping Based On Remote Sensing In The Central Valley California

Author : Han Liu
Publisher :
ISBN 13 :
Total Pages : 0 pages
Book Rating : 4.5/5 (825 download)

DOWNLOAD NOW!

Download or read book Understanding Spatiotemporal Variabilities in California Rangeland Forage Production Using Multi-source Remote Sensing Observations written by Han Liu and published by . This book was released on 2020 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Rangelands comprise more than 30% of global land area and provide a broad range of ecosystem services ‒ including food and forage production, soil and water resource protection, biodiversity, and wildlife habitat. In California, rangelands span about 57 million hectares, taking up nearly 60% of the state's land area, encompassing the Central Valley, as well as the coastal and foothill ranges. Forage production from these rangelands not only provides bases for many ecosystem services but also supports a 3-billion-dollar beef production every year. However, forage production in California has considerable year-to-year variation due to precipitation and temperature variability. The forage production also varies significantly from place to place due to California's complex topography-induced microclimates and spatial variability in soil properties. These spatiotemporal variations in rangeland productivity make balancing economic (grazing), conservation (habitat), and environmental (erosion/water quality) objectives a big challenge, especially during an era of climate change. The present work aims to examine relationships between rangeland forage production and its potential drivers at different scales to generate knowledge and tools that can help the ranching community and policymakers to make better management decisions. The first chapter focused on developing optimized methods to map forage production at a 30-cm resolution using multispectral remote sensing images. We estimated aboveground net primary production as a product of absorbed photosynthetically active radiation (APAR) derived from NDVI and light use efficiency (LUE), optimized as a function of topography and climate stressors. The estimated forage production agreed well with field measurements having an R2 of 0.80 and RMSE of 542 kg/ha. Cumulative NDVI and APAR were less correlated with measured biomass (R2= 0.68). This chapter demonstrated the utility of aerial and satellite remote sensing technology in supporting adaptive rangeland management by providing spatially explicit and near-real-time forage production estimates. Chapter two examined how climate and environmental factors regulate annual grassland forage production spatially across the state and among four ecoregions using machine learning models. We estimated annual forage production at a 30-m resolution over a 14-year period (2004-2017) using satellite images and data fusion techniques. The estimated forage production map was later fed into a gradient boosted regression trees (GBRT) model to quantify its relationship with its potential drivers. The GBRT model with 11 feature variables explained 67% of the variability in forage production across the state. Precipitation amount, especially in November (germination) and April (rapid growth), was found as the dominant driver for spatial variation in forage production, especially in drier ecoregions and during drier years. Seasonal distribution of precipitation and minimum air temperature showed a relatively stronger control on forage production in wetter regions and during wet years. Additionally, solar energy became more important in wetter ecoregions. This chapter demonstrated a cost-effective approach for rapid mapping and assessment of annual forage production with the potential for near real-time application. In chapter three, we investigated how climate regulates the temporal dynamics of annual gross primary production GPP in California's rangeland across geographical scales, based on 19 years of satellite record derived GPP at 500-meter resolution since 2001. We built GBRT models for each of 23 ecoregion subsections in our study area to relate annual GPP with 30 climatic variables and disentangled the partial dependence of GPP on each climate variable. Our analysis showed that GPP was most sensitive to growing season precipitation amount. We also found that years with precipitation more evenly distributed during the growing season had higher GPP. Warmer minimum air temperature in Winter was found to enhance the productivity in approximately 2/3rd of the subsections. In contrast, the average growing season mean air temperature and maximum air temperature showed a negative relationship with annual GPP. Across all rangelands, our models projected a slight decrease in GPP both by mid-century (2041-2059) and by the end of the century (2081-2099), when forced by downscaled future climate projection under Representative Concentration Pathway (RCP) 4.5 from four climate models. Under RCP 8.5, predicted GPP is going to increase in larger extents of the study area than under the intermediate scenario (RCP4.5), but the expected changes in GPP values were larger. Our machine learning based analysis highlighted the regional difference in GPP vulnerability to climate and provided insights on the intertwining and potentially counteracting effects of seasonal temperature and precipitation regimes.