Understand the spatial distribution of ground subsidence, and investigate the possibility of continual monitoring of ground subsidence, through the use of SAR interferometry

Target region: Northern area of the Kanto Plain (In the vicinity of the prefectural boundaries among Gunma, Tochigi, Saitama, and Ibaraki Prefectures).

This region is recognized as an area exhibiting a remarkable amount of ground subsidence caused by pumping of ground water for agricultural uses.

Data used: JERS-1 (Fuyô No. 1)/SAR level 0 data.

Data acquisition period: 1992 - 1998.

SAR interferometry analysis method: Two pass method using 50m interval digital elevation data (digital map) of the Geographical Survey Institute.

We obtained 20 interferograms from various time periods from 1992 to 1998. Based on these interferograms we created an isoline map showing the amount of vertical rise and fall movement.

In all of the cases the movement areas obtained from the SAR interferometry were in good agreement with the movement areas obtained from leveling, and the amount of the changes is also consistent. Here we show the result of the pair of October 2, 1993 and February 15, 1997 as a typical example.

Next, in order to evaluate the accuracy of the subsidence detected by SAR interferometry, we calculated for each analyzed interferometric pair the vertical movement at bench mark positions and compared these with the vertical movement obtained from the annual leveling survey. As the result, a strong positive correlation was seen between both measurements for all of the interferometric pairs.

We then performed a linear regression for the amount of the movement based on SAR interferometry and the amount of the movement determined from leveling. It shows a positive y intercept of about 3mm-3cm for each pair.

This means that there is a positive bias of about 3cm or less in the amount of ground subsidence obtained from SAR interferometry compared with leveling. (The figure shows the results for October 2, 1993 and February 15, 1997.) We think the difference between the results from SAR interferometry and leveling may be caused as follows: During SAR interferometry process, a relative position between two observations is required. However, the phase difference caused by ground subsidence makes the estimation of the relative position inaccurate. As a result, the correction of the topography-based phase difference is biased.

We concluded that even with the present level of technology it is possible to adequately detect ground subsidence of several cm in plains regions by using SAR interferometry. Further investigation is necessary, however, to determine why there is a difference in the absolute amount of the detected movement between the results obtained from SAR interferometry and levelings.

Although SAR interferometry does not achieve the level of accuracy provided by leveling, it does enable researchers to obtain movement over a wide area simultaneously. Moreover, if the satellite observation conditions are good, SAR interferometry has the positive feature of being able to make multiple observations in a year at a much lower measurement cost compared to leveling.

Accordingly, in the future SAR interferometry should be applied in ways that supplement , not replace, leveling. One way in which this might be applied, for example, would be to regularly monitor a region where ground subsidence is likely by using SAR interferometry, and carrying out concentrated leveling surveys when large ground subsidence is seen from the SAR interferometry results.