Landscape metric as indicators of disturbances in a beach-dune system.

ANA FAGGI; JOSÉ DADON; PATRICIA PERELMAN; GABRIELA POSSE

Hong Kong - China

Congreso; 15th International Conference on Environmental Bioindicators; 2007

the International Society of Environmental Bioindicators

In Argentina, more than 45% of the population is settled down within 100 km of the coast exerting big influence on the landscape. Their correct understanding is of concern, especially when addressing land use planning and conservation. The oceanic coast extends along 4725 km but only 800 km are sandy beaches and dunes, where vacation resorts are located. In our study we analyse changes in the landscape structure of the Atlantic beach-dune system caused by urbanisation in order to identify patterns of changes that can be incorporated into model describing coastal resorts development. Our hypothesis is that landscape metrics can be used as indicators for natural and human induced disturbances in a beach-dune system. The study area comprises 6 sectors along the Atlantic coast of The Buenos Aires province along a gradient of urbanisation defined by the proportion of the urbanised area in each sector. A land cover map for each sector was obtained using the spectral information provided by Landsat ETM+images. We performed an unsupervised classification with 20 classes with red, near red and mid infrared (spatial resolution 30 m). These classifications maps were grouped into 5 new classes: sandy beaches, vegetated dunes, forestations, dense urbanisation and disperse urbanisation. For the landscape structure analysis we used: INUMP: Index of number of patches, MPS mean patch size in hectares, AWMSI: Area weighted mean patch index as an average match adjusted perimeter area ratio. It gives an idea of the shape complexity. AWMPFD Area weighted mean patch fractal dimension to measure shape complexity. If the shape perimeters are simple this dimension approaches 1. MNN Mean nearest neighbour in meters as a measure of the open space between patches and as estimation of patch isolation. Number and form of dune patches can be linked to man pressure. In more populated resorts the dunes show superior number of patches of smaller size. This effect is coincident with decreasing values of AWMSI, indicative of a simplification of patch shape. This means smaller (9.71– 31,15 ha) and more isodiametric patches. In some resorts the deleterious effect of urbanisation on vegetated dunes is compensated by the progradation of the dunebeach system. At both extremes of the urbanisation gradient, patches show higher values of shape complexity. These could be associated with the process of dunes building which is wind dependent. By the urban areas the utilised metrics explain settlement differences among resorts. The disperse urbanisation show higher values of AWMSI in settlements with leap frog town growth. The indicator MNN pointed out to more isolated disperse urbanisation. Landscape metrics were not sensitive enough to describe differences by the classes sandy beach and forestation. The first is very much dependent on wind dynamics, a geomorphologic process essential in maintaining the landscape in its dynamic phase. Forestations features depend on the plantation actions carried on by man in each resort