The measurement of flood-stage discharge in mountain rivers is an enormous logistical challenge, particularly on large space and time scales, and the inevitable paucity of such data has impeded progress in our understanding of mountain landscape processes. We propose a novel solution that exploits new high resolution satellite imagery (2m, FORMOSAT-2) with unusually high revisit frequency (daily) over a region of very broad interest to the earth science community (Taiwan). Our proposed method entails the collation of a time series (2004-2005) of orthorecti*ed, georeferenced images of selected mountain river channels, in tandem with contemporaneous stage measurements of discharge at the nearest gauging station downstream. Patterns of inundation along selected channel reaches will be mapped in each image. With the aid of cross-channel profiles mapped in the field at low stage, we will make estimates of low width, depth, mean bed shear stress and mean low velocity. Once a broad range of discharges has been processed, we expect to be able to construct rating curves for these low measures as a function of the directly observed, at-station stage data. Combination of the rating models with the multi-decadal archive of station data in Taiwan will permit an assessment of the long-term probability distributions of low widths, depths and shear stresses along and across the selected bedrock channels. Measurements of this kind, albeit indirect, will be of enormous interest to the geomorphology community. However, the method is by no means guaranteed, because remote river gauging has never been attempted in mountain catchments; it remains to be seen whether the precision of the anticipated rating curves will suffice for rivers akin to those of the Taiwan Central Range. The main task of the pilot project will be to address this issue and to lay the foundation for a more comprehensive, longer-term assessment of Taiwanese mountain river discharge. Broader outcomes: While low measurements in steep mountain rivers are particularly sought after by those interested in bedrock channel processes, they are also of great interest beyond: from modelers of landscape evolution and the interaction between climate and tectonics, to hydrologists preoccupied with the assessment of flood hazard in populated mountain environments. Moreover, if the pilot project proves successful it will provide the Taiwanese space agency with an excellent example of the value of FORMOSAT-2 in natural hazard assessment, one of its principal missions. As a result, more attention will be paid to the regular acquisition of FORMOSAT-2 imagery in tropical mountain areas, an essential step if the elusive goal of a high-resolution, multitemporal image archive is to be achieved - at present, images of highland areas are a relatively low prioity (for FORMOSAT-2 and other equivalent satellites) and are only acquired daily during emergency periods (e.g. typhoons). Such an archive would be widely appreciated across the discplines of earth sciences, hydrology, ecology, not to mention in remote-sensing R&D itself. A-1
