Great forestry ecological engineerings, i.e. Conversion of Farmland to Forest (CFF) and Natural Forest Conservation (NFC), have played positive roles in promoting the ecological environment constructions in western China. The restoration of vegetation in these projects changed not only the composition but also the configuration of landscape pattern. In this paper, the landscape pattern of Hongya Country in 1994 and 2009 were compared to examine the effects of vegetation restoration on landscape pattern in the mountain-hilly region of the edge of Sichuan Basin, China. Landscape patterns at the two periods were both derived from Landsat TM images using a supervised classification procedure. The area of vegetation types were compared to examine the spatial distribution changes at different elevation and slope zones, and six landscape pattern indices were calculated to quantify the changes of landscape pattern at both landscape and class level. The elevation and slope zones were classified with the step sizes of 500m and 5° respectively, and the indices were calculated in Fragstats 3.4.
From 1994 to 2009, the total area of forest in Hongya increased by 5.72% from 60.58% to 66.30%, and the farmland and grassland decreased by 5.12% and 1.51% respectively. The increase of forestland and decrease of farmland and grassland are the major effects of vegetation restoration. Most of the conversions from farmland to forestland occurred under the elevation of 1000m, indicating a remarkable promotion of CFF to the change of landscape composition. Furthermore, as the dominant type, forest occupies more than 80% of total area in the region higher than 1500m. The vegetation restoration here was mainly influenced by the NFC and natural conditions, exhibiting little change of amount, but more change of landscape configuration. In region between the elevations of 1000m and 1500m, the amounts of both forestland and farmland were relatively stable, but remained some farm-forest conversions, and the landscape changes here were more complex, partly for the overlapping influences of CFF and NFC. At the slope aspect, forest in the region ≤ 5° increased obviously for the intense impact of human activities; moreover, in the region of near 25° forest also exhibited a little increasing trend, indicating more potential conversions from farmland to forest in the implement of CFF.
The overall landscape pattern change of Hongya exhibited a little defragmentation trend from 1994 to 2009. For farmland, influenced by CFF, patches were cut apart and converted to forest, resulted in more patches and smaller mean patch size, and the patches were more fragmented. While, forest was influenced by both the CFF and NFC, the gaps between forest patches were more filled, resulted in fewer patches and larger mean patch size, and the patches were more expanded and aggregated. In summary, as the consequences of vegetation restoration, forest was more continuously expanded to areas with lower elevation; while farmland was more shrunk, and its pattern was more fragmented.
At last, different change processes can result in different composition and configuration changes of landscape pattern. The method of traditional landscape pattern analysis cannot distinguish various processes of landscape change, in which the landscape and class level indices are indistinct and uninterpretable. There is a need to develop novel approaches to analysis the changes of landscape pattern for various change processes, especially for the interpretation of landscape indices.