Soil Moisture And Remotely Sensed Spectral Data In A Partial Canopy Cotton Field At The Maricopa Agricultural Center Pinal County Arizona 1988

Author : Matthew Romanko
Publisher :
ISBN 13 :
Total Pages : 139 pages
Book Rating : 4.:/5 (986 download)

DOWNLOAD NOW!

Download or read book Remote Sensing in Precision Agriculture written by Matthew Romanko and published by . This book was released on 2017 with total page 139 pages. Available in PDF, EPUB and Kindle. Book excerpt: Precision agricultural practices attempt to increase the efficiency of agricultural chemical usage in order to reduce pollution by implementing a variety of technology driven strategies. This study evaluated the effectiveness of remote sensing (RS) technology to model several important agricultural chemistry parameters. In situ hyperspectral and satellite multispectral measurements were used to examine the capability of a large range of soil and vegetation spectral indices (established indices from literature and spectral ratios) to model soil and vegetation chemistry, in particular plant chlorophyll and soil ammonia, nitrate, and phosphate contents. Data were collected from two farm fields (site 1: corn; and site 2: soybeans) in Wyandot County, Ohio at multiple times in 2015 (T1; early-May, through T6; early-November) including data collected prior to planting and fertilization and data collected post-harvest. The method used in the current study included the scaling up process from the in situ hyperspectral data to satellite observations (Landsat 8 OLI and Pleiades 1B) using aggregation of a) individual features reflectance measurements (soil or plants), and b) mixed reflectance data (soil and plants) to offer insight into the spatial, spectral, and temporal aspects of RS analyses. Laboratory testing for soil ammonium, phosphate and nitrate, and field measurements for plant chlorophyll were used in the assessment of the spectral indices. This research found that hyperspectral ratios were most effective for modeling the soil and plant parameters. The highest r2 values were reached for soil ammonia (r2 = 0.68); soil nitrate (r2 = 0.59) and soil phosphate (r2 = 0.75) using soil spectral ratios R1357/2024.3, R1379.5/2024.3, and R1854.7/1892.6, respectively (e.g., corn, site 1) during the peak of the growing season (T4). In addition to soil hyperspectral ratios, soil chemistry parameters were also statistically significantly correlated with several plant hyperspectral reflectance ratios, typically during early season (T2). Multispectral satellite data from Landsat 8 and Pleiades 1B showed that the relatively finer spectral resolution of the Landsat 8 sensor offered a distinct advantage in modeling the soil chemistry parameters when compared with Pleiades images during mid-season (T4). Landsat 8 was able to model each soil chemistry parameter more effectively producing the highest r2 values for site 1 soil ammonia (r2 = 0.31), site 2 soil nitrate (r2 = 0.33), site 1 soil phosphate (r2 = 0.41), and site 2 plant chlorophyll (r2 = 0.40) using satellite multispectral ratios; LS8 Rb1/b7, LS8 Rb2/b5, LS8 Rb3/b1, and; LS8 Rb4/b2 respectively. Analysis of spectrally aggregated hyperspectral data and satellite data indicated that Landsat 8 spectral bands in the Ultra-blue and SWIR regions were useful in RS estimation of soil chemistry from soil/plant/and mixed reflectance data. A variety of parameters known to influence soil and plant chemistry were also examined in this study. Soil moisture, pH, and potential wetness via Compound Topographic Index (CTI) index were statistically evaluated for influence on soil and plant chemistry measurements. This study showed that wetness significantly impacted the distribution of some soil chemistry parameters at times throughout the crop production season. The variability in canopy cover (referred to as `gap fraction' in this study) was also estimated at each sample location during the field campaign using Digital Hemispherical Photography (DHP) to combine hyperspectral soil and plant reflectance measurements in the appropriate proportions for comparison with satellite data. It was found that gap fraction explained 89% of the variation in Pleiades Rb2/b4 at site 1 and 51% of the variation in Pleiades Rb3/b1 at site 2.