Poster Session 3
Soil Moisture Determination Under Different Field conditions Using a Scatterometer and Space Borne SAR Systems
Shakil A. Romshoo, T. Nakaegawa, Masahiro Koike and Katumi Musaike
Mushiake lab, IIS, University of Tokyo, 7-22-1, Roppongi, Minato-ku, Tokyo-106,
Tel: (81)-3-3402-6231 ex. (2527), Fax: (81)-3-3402-2597
Soil moisture is an environmental descriptor that that integrates much of the land surface hydrology and is the interface for interaction between the solid Earth surface and life. It is a difficult variable to measure, not at one point in time, but in a consistent and spatially comprehensive basis. Much of out failure to translate the under-standing of point measurements to natural land-scapes c n be traced to a realization that soil moisture varies greatly in space with no obvious means to measure it. Remote sensing radar imaging systems with their all weather capability are a promising tool for data acquisition on a regular basic. For soil moisture, recent advances in remote sensing technology have demonstrated that soil moisture can be measured by a variety of techniques (Engman and Chauhan, 1995). How ever, only microwave technology has the ability to quantitatively measure soil moisture under a variety of topographic and vegetation cover conditions, so potentially it could be extended to routine measurements from a satellite system (Eng-man and Chauhan, 1995).
A number of experiments using sensors mounted in trucks, aircrafts, and space-crafts have shown that the moisture within a thin layer of soil on the order of 5 cm, can be accurately measured for bare soil and thinly vegetated surfaces. Earlier work by Ulaby (1974), using a ground Scatterometer, suggested that a C-band radar operating angles between
10o and 20o would be optimal for soil moisture extraction in the top 0.5 cm soil layer as the roughness effects could be minimized. Using a C-band Scatterometer measurement of a tall-grass prairie, Martin et al. (1989) observed that viewing angles Martin et al. (1989) observed that viewing angles of 30o and 45o also produced strong correlations with soil moisture. The capability to measure soil moisture from a space platform was also demonstrated qualitatively with SEASAT SAR data (Blanchard and Chang, 1983) and quantitatively through experiments with shuttle Imaging Radar SIR-B, carrying a L-band SAR system (Dobson and Ulaby, 1986a)
The objective of this research has been to explore the possibilities of temporal analysis of soil moisture change detection under different conditions using the ground based and space borne Synthetic Aperture Radar (SAR) systems and also to analyze and understand the influences of the vegetation on the s0 sensitivity to the soil moisture detection. For temporal analysis of soil moisture using radar images, it is desirable to examine a period over which marginal changes in roughness and vegetation are expected.
The study has been conducted in different fields under varied surface conditions. Three fields located at different areas in Chiba prefecture, Japan viz., Kemigawa, located at
35.67o north latitude and 140.08o east longitude, Futawa., located at
35.30o north latitude and 139.31o east longitude and Inbanuma were chosen for the study. The former two, comprising a number of play grounds, each were selected for the field campaigns during the RADARSAT over pass over these areas while as the later located at
35.74o north latitude and 140.20o east longitude was selected for the soil moisture determination under paddy field conditions. The dimensions of the Kemigawa and Nihon fields were 17.76 ha and 3.56 ha respectively while as that of paddy field were 100m *200m. The topography of these fields is quite flat. The soil types of the fields are Narita soils. The two fields chosen for RADARSAT campaign have a composite land use of both bare and grass areas which did not change much during the campaign period. The Nihon field was comparatively roughter than Kemigawa fields.
The Scatterometer employed in collecting the data in this study is the 5.2 GHz microwave active Scatterometer (5.8 cm wavelength). This is a mobile van-mounted system capable of making scattering measurement at a single frequency (C-Band) and employs a dual antenna system con-figured to allow all the linear polarization modes of operation. Measurements can be made at angles of incidence from zero degrees (nadir) to 70degrees. The details of the system specifications and characteristics can be found from Musiake et al., (1987). During the experiments coinciding with the RADARSAT revisits, The measurements were recorded at HH polarization and
23o and 35o incidence angles compared to the RADARSAT sensor configuration of HH polarization and
21o-27o angles of incidence in the standard beam mode (SI). In case of paddy field, the measurements were also recorded at 43oincidence angle in addition to the other two angles. The choice of the active microwave radar sensors on board the space platforms like ERS and JERS satellite systems.
Fields Data Collection
The field campaign for the RADARSAT in two fields viz kemigawa and Nihon Univ. was conducted six times with the first one starting on 15 May, 97 and the last one conducted on 17 Dec., 97. In addition to measuring the volumetric soil moisture and Time domain reflectometry (TDR) during these campaigns, the C-band Scatterometer was also used during the campaign. In case of soil moisture determination under paddy crop, besides measuring the surface characteristics like soil moisture by volumetric and TDR means, rms height, the plant parameters like LAI, density and height, of plants were also recorded in the field. The paddy crop was raised in this field by planting the seeds in rows rather than the transplanting the seeds in rows rather than the transplanting of the seedlings which is a common practice all over the globe. The tillage practice in the area is concerned. The gaps in the time history of data are either due to system faults or due to the flooded field conditions.
Satellite Data-Sets Used
In all, six RADARSAT scenes of standard beam mode were used for two sites i.e. Kemigawa and a C-band SAR system with a frequency of 5.4 GHz and HH polarization with incidence angle ranging from
21o – 27o across the swath of the standard beam (SI) images. The RADARSAT data for the following days of 1997. was used: May 5, 15, May 18, July 16, August 9 and August 19. Each scene covers a minimum ground swath of 100 Km square. The nominal spatial resolution in the range direction is 26m for the S1 mode at near range. The azimuth resolution is 27 m. the product has pixel dimensions of 12.5 m and are a result of processing of four independent azimuth samples, or four independent azimuth samples, or four looks, in the along track antenna beam dwell time and then non-coherently summing of the coherent speckle noise in the image to provide improved radiometric resolution.
Table 2: Summary of correlation coefficients (r) and slope of so for Scatterometer and RADARSAT at Kemigawa and Nihon Univ. fields.