Comparison of jers-1 and radarsat synthetic aperture
Radar data for mapping mangrove and its biomass
Biomass estimation
The stepwise regression analysis indicated that mangrove biomass in both image can
reasonably be estimated by:
JERS-1 SAR, B = 92.431s° + 1381.5 (1)
Radarsat SAR, B = 1004. 7e0.1352x (2)
where
B = total biomass in ton/ha.;
s° = radar backscatter coefficient derived using;
20 log (DN) – 68.5 for JER-1 SAR and 10 log (DN
2 /A) + 10 log sin I for Radarsat
SAR;
A = scaling gain (5695770.5);
I = Incident angle (20.2°);
DN=digital number recorded from images.
The computed biomass using the relationship is shown in equation (1) and equation
(2) and is given in Figure 3. These computed biomass are then compared with
biomass derived using most recent record of tree-age of the area compiled during
fieldwork on 1998. For accuracy assessment, biomass value was divided to seven
clasess in 100 ton/ha. range. Using random generation or more than 100 samples, the
overall average accuracy of computed biomass in the seven tonnage categories is
only at 40 percent. These results confirmed to recent similar biomass studies of
mangrove forest using SAR that was carried in French Guiana and Bangladesh
respectively. (Mougin et al., 1999). Detailed producer’s and user’s accuracy
information is given in Table 3.
Table 3: Accuracy assessment of biomass estimation statistics for JER-1 SAR and Radarsat.
| Data | JERS-1 SAR | Radarsat |
| Biomass (ton/ha.) | User’s Accuracy | Producer’s Accuracy | User’s Accuracy | Producer’s Accuracy |
Less than 100 | 50 % | 100% | 75 % | 75 % |
| 101 – 200 | 29 % | 71% | 29 % | 46 % |
| 201 – 300 | 29% | 42% | 33 % | 36 % |
| 301 – 400 | 50% | 42% | 31 % | 35 % |
| 401 – 500 | 45% | 36% | 45 % | 38 % |
| 501 – 600 | 12.5% | 27% | 31 % | 29 % |
| Over than 601 | 60% | 27% | 50 % | 33 % |
| Overall Accuracy | 39.3 | 36.8 |
| KHAT | 27.2 | 24.4 |
(a)
(b)
Figure 3:
Biomass derived from (a) JERS-1 SAR, and (b)Radarsat backscatter
An interesting result to be noted is that biomass for less than 200 ton/ha can be
determined more accurately using SAR: For biomass less than 100 ton/ha were
derived perfectly using the model adopted, and 71 percent accuracy is reported for
biomass in the range of 100-200 ton/ha. The accuracy then degrades to 27 percent for
biomass of more than 600 ton/ha. Lower accuracy was observed as biomass increases
– two reasons that might contribute to this accuracy trend are: (1) non-representative
regression model due to limited samples used in generating the biomass-backscatter
relationship, (2) mixed species in area of larger biomass but only dominating
pioneering species were accounted in derived biomass. However both these factors
are yet to be improvised in near future due to the restriction in obtaining logistic
support on comprehensive samples.
Conclusion
The results demonstrates the utility of SAR data as potential source in mapping
classes and indicator for biomass. Although there has been limited availability of
exhaustive sampling points particularly on focussed mangrove forest, but the results
indicates the evidence of C band and L band utility for mangrove mapping and
biomass estimation. The mangrove biomass estimation was found related to JERS-1
and Radarsat backscatter coefficient at r
2 = 0.5 and 0.31. The on-going and future task
of this study is for decomposed forest’s SAR backscatter element to biomass
estimation to other forest types.
References
- Beaudoin, A., T. LeTeon, S. Goze, E. Nezry, A. Lopes, E. Mougin, C. C. Hsu, H. C. Han, J.
A. Kong and R. T. Shin (1994). “Retrieval of forest biomass from SAR data.” Int. J.
Remote Sensing. 15 ; 1117-1124.
- Clough, B. F. (1993). The status and value of mangrove forests in Indonesia, Malaysia and
Thailand: Summary. The economic and environmental values of mangrove forests and their
present state of conservation in the South-East Asia/Pacific Region. P 1-10. Institute of
Marine Science. Camberra, Australia.
-
Imhoff, M. L. (1995). “A theoretical analysis of the effect of forest structure on synthetic
aperture radar backscatter and the remote sensing of biomass.” IEEE Trans. Geosci. Remote
Sensing .33(2): 341-352
-
Lopes A., R. Touzi and E. Nezry (1990) “Adaptive filters and Scene Heterogeneity.” IEEE
Trans. Geosci. Remote Sensing. 28(6); 992- 1000.
-
Mazlan Hashim, Wan Hazli Wan Kadir, Lee Ken Yoong (1999). Global Rainforest Mapping
Activities in Malaysia: Radar Remote Sensing for Forest Survey and Biomass Indicators;
JERS-1 Science Program 99 PI Reports: Global Forest Monitoring and SAR Interferometry,
Earth Observation Research Centre, National Space Development of Japan, 63-70.
-
Mougin, E., C. Proisy and G. Marty (1999) “Multifrequency and multipolarization radar
backscattering from mangrove forest.” Submitted to IEEE Transsactions on Geoscience and
Remote Sensing.
-
Paudyal, D. R and J. Aschbacter (1993) “Evaluation and performance test of selected SAR
speckles filter.” Presented at the International Symposium “Operationalization Org. Remote
Sensing” ITC Enschede, The Netherlands
