Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal RES1369

TitleEstimation of crops biomass using TerraSAR-X data in Navarre, Spain
Investigator Alvarez-Mozos, Jesus - Public University of Navarre, Projects and Rural Engineering
Team Member
Dr Arregui, Luis M. - Public University of Navarre, Agricultural Production
Mrs Larrañaga, Arantzazu - Public University of Navarre, Projects and Rural Engineering
Prof. Gonzalez-Audicana, Maria - Public University of Navarre, Projects and Rural Engineering
SummaryA. Objectives The main objective is to evaluate the sensitivity of different types of SAR observations to crops biomass. Hence, we plan to evaluate multi-temporal, multi-incidence angle, multi-polarisation and multi-frequency data. B. Methods In the literature several approaches for SAR based crop biomass estimation can be found. Different techniques have been developed and applied depending on the characteristics of the available SAR data. The analysis will focus on cereal crops where the main interest of biomass estimations is related to their use as fuel for biomass energy plants. Ground truth data on crops biomass along the season will be collected coinciding with image acquisitions on 20 control fields. The different types of SAR data to be used in this project can be separated according to their band (C and X-bands) and polarization (single-pol, dual-pol, quad-pol). The selected approach to obtain biomass estimates will depend on the polarization of the used SAR data: - Single-pol data The backscatter coefficient (s0can) observed over cereal canopies depends on the dielectric and geometric properties of both the soil surface and the vegetation canopy. Several models have been proposed to calculate s0can once the required soil and vegetation parameters are known. From an operational point of view semi-empirical models are usually preferred to theoretical models. In particular, the semi-empirical Water Cloud Model (WCM) has been widely applied because of its simplicity and adequacy (Wigneron et al., 1999). - Dual-pol data Previous studies on narrow-leaf crops showed a strong relationship between crop parameters, such as biomass and LAI, and the co-polarized backscatter ratio s0HH/s0VV for C-band (Mattia et al., 2003). Compared to C- and L-band data relatively few studies have been carried out using X-band data acquired over agricultural areas. The vertical polarization channel (VV) will probably be positively correlated with crop growth status, since the vertical stems will likely contribute to the backscatter at this band. On the other hand, cross polarized observations (HV or VH) will probably reflect phenological changes that might cause depolarization. Therefore, at X-band VV/VH dual polarized observations will be preferred. - Quad-pol data Quad-pol, or polarimetric, observations provide a much more complete description of the observed surface. In the case of cereal crop canopies, there are a number of observables obtained from quad-pol observations that can be related to crop condition, and hence to biomass: backscatter ratios in different basis, phase differences, coherences, polarization decompositions. C. Data requirements The prefered TerraSAR-X products are Dual Polarised StripMap Products, with VV/VHpolarisations and incidence angles between 25º and 35º. The processing modeselected is Multi Look Ground Range Detected (MGD). This product is alreadyprojected in ground range hand has a reduced speckle. The scenes will be later geocodedusing a 5 m resolution DEM and control points in the area. The numberof scenes required is around ten, acquired every 15 days or so covering theagricultural season between 1/March and 1/August 2012. D. Deliverables The expected deliverables are the following: -Methodology for biomass estimation using single-pol radar data and accuracy analysis. -Methodology for biomass estimation using dual-pol radar data and accuracy analysis. -Methodology for biomass estimation using quad-pol radar data and accuracy analysis. -Recommendations on the required number of scenes and on the best image acquisition dates for each approach. -Influence of field and terrain characteristics (field size, type of cereal crop, slope and aspect) on the biomass estimation accuracy. -Analysis of Antecedent Precipitation Indexes related to soil moisture.

Back to list of proposals

© DLR 2004-2016