In recent years, Global Navigation Satellite Systems Reflectometry （GNSS-R） is developed to estimate soil moisture content （SMC） as a new remote sensing tool. Signal error of Global Positioning System （GPS） bist...In recent years, Global Navigation Satellite Systems Reflectometry （GNSS-R） is developed to estimate soil moisture content （SMC） as a new remote sensing tool. Signal error of Global Positioning System （GPS） bistatic radar is an important factor that affects the accuracy of SMC estimation. In this paper, two methods of GPS signal calibration involving both the direct and reflected signals are introduced, and a detailed explanation of the theoretical basis for such methods is given. An improved SMC estimation model utilizing calibrated GPS L-band signals is proposed, and the estimation accuracy is validated using the airborne GPS data from the Soil Moisture Experiment in 2002 （SMEX02）. We choose 21 sites with soybean and corn in the Walnut Creek region of the US for validation. The sites are divided into three categories according to their vegetation cover： bare soil, mid-vegetation cover （Mid-Veg）, and high-vegetation cover （High-Veg）. The accuracy of SMC estimation .is 11.17% for bare soil and 8.12% for Mid-Veg sites, much better than that of the traditional model. For High-Veg sites, the effect of signal attenuation due to vegetation cover is preliminarily taken into consideration and a linear model related to Normalized Difference Vegetation Indices （NDVI） is adopted to obtain a factor for rectifying the ＂over-calibration＂, and the error for High-Veg sites is finally reduced to 3.81%.展开更多
基金Supported by the National "12th Five-Year Plan" Pre-Research Program on Civil Space
文摘In recent years, Global Navigation Satellite Systems Reflectometry （GNSS-R） is developed to estimate soil moisture content （SMC） as a new remote sensing tool. Signal error of Global Positioning System （GPS） bistatic radar is an important factor that affects the accuracy of SMC estimation. In this paper, two methods of GPS signal calibration involving both the direct and reflected signals are introduced, and a detailed explanation of the theoretical basis for such methods is given. An improved SMC estimation model utilizing calibrated GPS L-band signals is proposed, and the estimation accuracy is validated using the airborne GPS data from the Soil Moisture Experiment in 2002 （SMEX02）. We choose 21 sites with soybean and corn in the Walnut Creek region of the US for validation. The sites are divided into three categories according to their vegetation cover： bare soil, mid-vegetation cover （Mid-Veg）, and high-vegetation cover （High-Veg）. The accuracy of SMC estimation .is 11.17% for bare soil and 8.12% for Mid-Veg sites, much better than that of the traditional model. For High-Veg sites, the effect of signal attenuation due to vegetation cover is preliminarily taken into consideration and a linear model related to Normalized Difference Vegetation Indices （NDVI） is adopted to obtain a factor for rectifying the ＂over-calibration＂, and the error for High-Veg sites is finally reduced to 3.81%.
