* Corresponding author. E-mail address: fbruce@internext.com.br (B.R. Forsberg). Quaternary International 72 (2000) 61}66 Tectonic fault control of wetland distributions in the Central Amazon revealed by JERS-1 radar imagery Bruce R. Forsberg*, Yasuaki Hashimoto,A s ke Rosenqvist, Fernando Pellon de Miranda Coordenac 7 a J o de Pesquisas em Ecologia, Instituto Nacional de Pesquisas da Amazonia, Caixa Postal 478, Manaus, AM, 69011-970, Brazil Analysis and Research Department, Remote Sensing Technology Center of Japan, Roppongi First Bldg. 8F, Roppongi 1-9-9, Minato-ku, Tokyo 106, Japan Joint Research Center of the European Commission Space Applications Institute - MTV Unit, TP 440, I-21020 ISPRA (VA), Italy Research and Development Center, Petrobras, Cidade Universita & ria Q.7 - Ilha do Funda J o, Rio de Janeiro, RJ 21949-900, Brazil Abstract A mosaic of JERS-1 L-band synthetic aperture radar (SAR) images was used to investigate the in#uence of tectonic faults on wetland distributions in &200,000 km of central Amazon lowland (0}43 Lat. S, 60}643 Long. W). The geographic distribution of #ooded wetland was clearly evident on the mosaic due to the unique characteristics of L-band radar. Two distinct linear boundaries were encountered limiting the northern distribution of wetlands, one north of the Negro river main channel and west of the Branco river, oriented WNW}ESE, and one south of the Negro main channel and east of the Branco, oriented NW}SE. The orientations and positions of these boundaries were consistent with the prevailing tectonic fracture pattern in the region. Geophysical, pedological and geomorphological data supported the hypothesis that these boundaries are tectonically controlled. The ecological, economic and biogeochemical implications of the observed wetland distribution were considered. The distributions of wetland-dependent biota were expected to be severely limited north of the Negro main channel. Fish production, wetland timber yields and methane emissions were also predicted to be exceptionally low in this region.  2000 Elsevier Science Ltd and INQUA. All rights reserved. 1. Introduction It has been estimated that 20% of the Amazon lowland is covered by permanently or seasonally #ooded wet- lands (Junk, 1993). River #oodplains are the dominant wetland habitat in the Amazon. Inter#uvial swamps and #ooded savannas also cover extensive areas in some regions. These #ooded environments provide important habitat for aquatic #ora and fauna (Junk, 1997) and play a key role in sustaining regional "sh production (Bayley, 1989; Forsberg et al., 1993). They are also a globally signi"cant source of methane and other trace gases essen- tial for climate regulation (Devol et al., 1994). Accurate information on wetland distributions is currently needed to improve estimates of habitat availability and calculate regional contributions of trace gases to the troposphere. The geographic distribution of most wetlands in the Amazon Basin is still poorly de"ned. The existing maps of #oodplain distributions are based primarily on visual interpretations of X-band radar images (Projeto RADAMBRASIL, 1972) which, while useful for distin- guishing many geomorphological features, cannot di!er- entiate between dry and #ooded vegetation due to the short (3.1 cm) X-band wavelength. L-band Synthetic Ap- erture Radar (SAR) imagery, derived from NASDA's JERS-1 satellite, now provides an accurate means of identifying and mapping wetland areas (Kasischke et al., 1997; Hess et al., 1998; Rosenqvist et al., 1998). The longer (23.5 cm) L-band radar e$ciently penetrates woody and herbaceous vegetation and produces a strong dihedral `double-bouncea return signal as it re#ects o! the water surface and tree trunks in #ooded habitats (Kasischke et al., 1997). This results in a characteristic bright area on a SAR image, easily distinguishable from the gray and black features produced by dry vegetation and water, respectively (see Fig. 1). A mosaic of 90 digi- tally assembled JERS-1 radar images, acquired from February 23}March 3, 1993 and covering approximately 400,000 km of the central Amazon basin, has provided the "rst precise large-scale data on wetland distributions in this region (Hashimoto and Tsuchiya, 1995). We used a portion of this mosaic here to investigate the e!ects of neotectonic activity on wetland distributions in the cen- tral Amazon lowlands. JQI=472=Shantha=Venkatachala=BG 1040-6182/00/$20.00  2000 Elsevier Science Ltd and INQUA. All rights reserved. PII: S 1 0 4 0 - 6 1 8 2 ( 0 0 ) 0 0 0 2 1 - 5