Calibration and Modelling of Vegetation Structure Parameter B for L-Meb Models Over Agricultural Sites

39 Pages Posted: 29 Oct 2024

See all articles by Zhang Hui

Zhang Hui

Zhejiang University

Hongquan Wang

Government of Canada - Agriculture and Agri-Food Canada

Zhibin Sun

Nanjing Normal University

Lun Gao

University of Illinois at Urbana-Champaign

Yixiang Wang

Zhejiang Agriculture and Forestry University

Cheng Tong

Hangzhou City University

Luyao Zhu

Zhejiang University

Zhou Shi

Zhejiang University

Multiple version iconThere are 2 versions of this paper

Abstract

Passive microwave observations of vegetation water content (VWC) are important for monitoring vegetation growth and aboveground biomass. To constrain the L-band Microwave Emission of the Biosphere (L-MEB) model for NASA’s Soil Moisture Active Passive (SMAP) mission, vegetation structure parameter b is used to estimate VWC fro m the climatological Normalized Difference Vegetation Index (NDVI). However, little is known about what and how the driving factors impact the parameter b and its modeling due to the lack of quantitative analysis. Using the Passive Active L-band Sensor (PALS) data from the SMAPVEX16-MB campaign, this study firstly calibrated parameter b for canola, wheat and soybean. Afterwards, a Random Forest algorithm was employed to model b-values using both dual-polarized observations with input features screened by correlation analysis and stepwise regression, the R2 of which exceeds 0.82. To enhance interpretability of the model, multivariable-linear regression was also employed, achieving an R2 exceeding 0.75. The results showed three crops had increasing trends in b-values with growth, except for wheat which exhibited a decreasing trend at v-polarization. Wheat and soybean indicated sharp changes in b-values particularly during late growth stages, while canola exhibited slower changes. The b-values were higher at v-polarization than h-polarization, except for wheat during early growth stage and at the end of late growth stages. The multivariable regression with vegetation type as dummy variables indicated that crop type significantly affected b-values, with canola and wheat positively impacting them at h-polarization but negatively at v-polarization. Additionally, key factors affecting b -values included VWC, biomass and LAI at h-polarization, while VWC and biomass dominated at v-polarization. Notably, VWC and brightness temperature at h-polarization (TBh) negatively impacted b-values at both polarizations. Overall, this study enhances our understanding of parameter b and the VWC retrieval and also contribute to model parameter b at the pixel scale of the satellite observations.

Keywords: L-MEB model, Parameter b, Passive microwave, SMAPVEX16

Suggested Citation

Hui, Zhang and Wang, Hongquan and Sun, Zhibin and Gao, Lun and Wang, Yixiang and Tong, Cheng and Zhu, Luyao and Shi, Zhou, Calibration and Modelling of Vegetation Structure Parameter B for L-Meb Models Over Agricultural Sites. Available at SSRN: https://ssrn.com/abstract=5003390 or http://dx.doi.org/10.2139/ssrn.5003390

Zhang Hui

Zhejiang University ( email )

Hongquan Wang

Government of Canada - Agriculture and Agri-Food Canada ( email )

Performance and Analysis Directorate
Policy Branch
Ottawa, K1V 0C6
Canada

Zhibin Sun

Nanjing Normal University ( email )

Ninghai Road 122, Gulou District
Nanjing, 210046
China

Lun Gao

University of Illinois at Urbana-Champaign ( email )

601 E John St
Champaign, IL 61820
United States

Yixiang Wang

Zhejiang Agriculture and Forestry University ( email )

Cheng Tong

Hangzhou City University ( email )

Hangzhou
China

Luyao Zhu

Zhejiang University ( email )

38 Zheda Road
Hangzhou, 310058
China

Zhou Shi (Contact Author)

Zhejiang University ( email )

38 Zheda Road
Hangzhou, 310058
China

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
23
Abstract Views
114
PlumX Metrics