The Relationship Between Watershed Landscape Pattern And Total Nitrogen Or Total Phosphorus In Dianchi Lake Sediments

Watershed landscape spatial pattern is very important for ecological security and sustainable development in a watershed. Pattern changes may affect landscape functions and ecosystem services. Dianchi is a typical plateau lake in Yunnan Province, and the topography of its watershed is quite unique. Generally, Dianchi lake watershed plays an irreplaceable role in the social economic development process in Yunnan. However, with the rapid economic development and the expansion of Kunming City, Dianchi lake pollution load is increasing, and the watershed landscape pattern changes dramatically. Therefore, it’s critical to study the relationship between landscape pattern and water quality in Dianchi lake watershed.This research based on landscape ecology theory, and it was proceed by using some software, such as ERDAS IMAGINE9.2and ArcGIS10.1. The satellite imageries used to derive landscape pattern changes were collected from the year of1957,1972,1992and2010, respectively. And the spatial scales were divided into three levels:①level1including Dianchi lake watershed;②level2including Dianchi lakeshore and upland;③level3just including Dianchi lakeshore. Then some landscape pattern indices (LPIs) were selected to describe the pattern changes at different scales. At the same time, conversion matrix of landscape types were carried out to know how pattern changes. Lastly, the geographically weighted regression (GWR) models were developed in ArcGIS10.1. In those GWR models, the dependent variables were total nitrogen and total phosphorus in Dianchi sediments, meanwhile independent variables were interspersion and juxtaposition index (IJI) and the percentage of landscape types (PLAND) at class level. So that would be clear for the relationship between landscape pattern and total nitrogen or phosphorus in Dianchi lake watershed. The main results are as follows:(1) From1992to2010, the landscape fragmentation was tend to worse at level1spatial scale. And during this time period at the same scale, the area changes of landscape types accounted for45.40%. That changes were mainly embodied the rapid increase of irrigated land, urban and other woodland, which relative change rate were578.01%，173.27%and132.89%, respectively. The increase in irrigated land and urban area mainly occurred in paddy and arid land in both. In addition, rural residential region was an extra major contribution type to urban increase. And the main contribution types to other woodland increase were shrub, forest and paddy.(2) During1957-1972and1992-2010, the landscape fragmentation in later was better than in the former at level3spatial scale. At the same scale, the area changes of landscape types accounted for11.74%during1957-1972, while it accounted for24.44%and42.82%during1972-1992and1992-2010, respectively.24.44%area changes were principally embodied the increase of reservoir ponds and urban, which relative change rate were160.65%and128.33%, respectively. Similarly,42.82%area changes were mostly due to the sharp increase of irrigated land, other woodland and urban, which relative change rate up to1236.77%,358.13%and257.22%. respectively. Overall, those increase mainly occurred in paddy and arid land.(3) There were obvious scale effect for the relationship between watershed landscape pattern and total nitrogen or total phosphorus in Dianchi lake sediments. For example:①at level3spatial scale, the IJI factors of paddy and arid land were negatively correlated with TN or TP in1972. However, they were positive correlation in2010.②The correlation between PLAND and TN at level3spatial scale was significantly better than at level1spatial scale.(4) Meanwhile, it also had some law in common for the relationship between watershed landscape pattern and total nitrogen or total phosphorus in Dianchi lake sediments. No matter which spatial and temporal scales,①the IJI of river, lake and urban were positively correlated with TN or TP.②The PLAND of lake were negatively correlated with TN or TP.③The PLAND of urban, residential region, transportation land, irrigated land and artificial woodland were positively correlated with TP.