Detecting trends in seagrass abundance using aerial photograph interpretation: Problems arising with the evolution of mapping methods |
| |
| Authors: | A J Meehan R J Williams F A Watford |
| |
| Institution: | 1. New South Wales Fisheries Office of Conservation, Post Office Box 21, 2230, Cronulla, NSW, Australia
|
| |
| Abstract: | In the late 1970s and early 1980s an inventory of seagrass distribution in southeast Australia was made from multiscale aerial photographs using a simple analog mapping method, referred to here as theCamera lucida technique. Four priority estuaries were remapped in 2003, also using multi-scale aerial photographs, but using a geographic information system (GIS)-based analog mapping technique. We quantified the degree of error that arose when estimates of seagrass distribution were compared for the two techniques be ressessing. with GIS, the same aerial photographs used earlier for four estuaries: Port Hacking, Bermagui River, Merimbula Lake, and St Georges Basin.Camera lucida produced a larger estimate of total seagrass area than GIS in all four locations: 8% of total seagrass area or 12.9 ha in Port Hacking, 15% or 5 ha in Bermagui River, 20% or 38.0 ha in Merimbula Lake, and 243% or 502.2 ha in St Georges Basin. The consistent and moderate discrepancy in Port Hacking, Bermagui River, and Merimbula Lake was attributed mainly to differences betweenCamera lucida and GIS, such as the tendency ofCamera lucida to overestimate seagrass area by amalgamating disjunct seagrass patches into continuous meadows. The large discrepancy at St Georges Basin was attributed to interoperator error; i.e., the operators who mapped photographs taken in 1979 had widely differing interpretations of the aerial photographs. A 20-yr trend was calculated for each estuary using either theCamera lucida or revised GIS value as the start point. Perceived management status of the seagrass depended greatly on the initial data used. In St Georges Basin, a catastrophic decline (65%, 553.9 ha) occurred based on theCamera lucida-GIS comparison but only a small decline (14.7%, 51.8 ha) when all photos were GIS assessed, In view of the technical developments in habitat mapping, evaluation of change in seagrass abundance using remote sensing needs an assessment of intra-operator error, inter-operator error, and systemic error before maps produced by superseded technology are considered useful baselines. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
