The Export Characteristics Of Nitrogen Phosphorus And Carbon Based On Landuse Structure And Landscape Pattern

With the development of urbanization in China, suburban agricultural area was expanding constantly and became a particular area with dual functions which were giving security of agricultural products and ecological environment for city. Due to the superiority of geographic location and agricultural infrastructure, the main industrial structure of the area were vegetable food cultivation and intensive livestock farming. However, non-point source pollution became significant to water environment pollution and threatens ecological environment seriously because of excessive use of fertilizers with low utilization ratio and lack of collection and treatment facilities for livestock wastewater, domestic sewage and industrial wastewater.This study was conducted in Jinjing watershed of Changsha city, Hunan province, in the suburban agriculture regions. Utilizing GIS and RS technology as a platform, continuous water sampling was carried out from Dec.2009to Dec.2011, and analysed the export characteristics of nitrogen phosphorus and carbon based on using multivariate statistical techniques, afterwards combined with the landuse structure and landscape pattern within the watershed to discuss the relationship between the export and the factors. And we hope it will provide scientific references for non-point source pollution control measures establishing and sustainable development of suburban agricultural area. The present study came to make important results and achievements as the following.(1) According to the data based on two year sampling, nitrogen pollution in the watershed was extremely serious, the average total N (TN) concentration of stream water was3.67mg·L-1. In the inter-annual scales N outflow in surface runoff was dominated by nitrate N (NO3--N) and ammonium N (NH4+-N) that was better, then the Pearson correlation between NH4+-N and TN achieved extremely significant positive, and for NO3--N and nitrite N (NO2--N) the Pearson correlation reached significant positive, respectively. But in the season scales N outflow in surface runoff was dominated by NO3--N except winter. The seasonal variation for the export concentration of TN and NH4+-N was distinct in the following order:WI> SP> FA> SU, compared with the NO3--N output in the following order:SP> WI, FA> SU. The concentration and temporal change of different form of nitrogen was extremely different in different sample sites. However, the total phosphorus (TP) concentration of stream water during the monitoring period was low ranged from0.03to0.26mg·L-1except NO.1sample sites with a vague seasonal variation. According to the data based on half year sampling, the average dissolved organic carbon concentration (DOC) of stream water was4.70mg·L-1, and6.47mg·L-1for dissolved inorganic carbon (DIC). In the monitoring period, the concentration of DOC descended first, and the minimum value was1.26mg·L-1in September, and then the concentration ascended, but the maximum value was7.55mg·L-1in July. Nevertheless, the concentration of DIC ascended continually ranged from1.5to11.5mg-L"1. To evaluating eutrophication of the stream water, eutrophication index (EI) was counting by a universal index formula a logarithmic power function, and the mean of El value was52.2, which represent the grade of eutrophication was "rich". The spatial variation of El value in the watershed was remarkable. In the monitoring period, El value descended first, and the minimum value was45.6in September, and then the concentration ascended, but the maximum value was58.3in November, but the grade of eutrophication was " rich " all the time.(2) Hierarchical cluster analysis grouped20sample sites into three clusters, relatively highly polluted (HP), medium polluted (MP) and less polluted (LP) sites, based on the similarity of water quality characteristics. The following order:HP> MP> LP was observed in the export of NH4+-N, TN, TP, DOC and DIC, but the order for NO3--N was MP> HP> LP. Factor analysis, applied to the date sets of the three different groups obtained from cluster analysis, resulted in three, three and four latent factors explaining90.75,90.78and89.02%of the total variance in water quality date sets of HP, MP and LP areas, respectively. The varifactors obtained from factor analysis indicated that HP area sites was mostly influenced by point source pollution (domestic wastewater, livestock wastewater and eutrophic lake water), and the pollution rank of HP area sites were as follows:S8> S1> S3> S5> S20>S16. Likewise, MP area sites was mostly influenced by agricultural non-point source pollution (fertilizer, water and soil erosion), and the pollution rank were as follows:S19> S4> S7> S6> S15. And in LP area sites the water and soil erosion was the main factor, and the pollution rank were as follows:S17> S18> S14> S10> S13> S12> S9> S2>S11.(3) In the basin scale, the Jinjing river watershed was dominated by forest land, and there were significant differences between landuse structure in each sub-basin. The major landuse of HP area was cultivated land, compared with LP areas was forest land, and in MP areas the ratio of forest and cultivated land were equal. Considering the landuse from elevation, slope and buffer, the forest land was distributing in middle highlands with slope more than15°and the distance to river was between300m to700m; the cultivated land and habitation land were distributing in low lands with slope less than5°and the distance was distinct in different sub-basin; the water body was locating in highlands with slope less than5°and the distance was father.Without HP area sites, analyzing the source or sink feather of landuse type was performing in different time scale, it was all concluded that cultivated land and habitation land exported significant sources of nitrogen, phosphorus and carbon, while forest and water body could be considered as nitrogen sinks. Multiple stepwise regression analysis showed in the inter-annual scales paddy field played a crucial factor in spatial variation of TN and NO3--N, compared with forest land for NH4+-N and habitation land for DIC. But in the seasonal scales paddy field also played a crucial factor in spatial variation of TN except winter whose major factor was forest land, and paddy field mostly influenced the spatial variation of NO3--N in summer contrast with water body in fall; forest land had the most influence on NH4+-N in winter and summer compared to paddy field for spring; habitation land and forest land had the most influence on DIC separately in summer and fall. No regression equations were found between land use structure and DOC or TP.(4) In landscape scale we selected12landscape metrics to establishing landscape pattern for Jinjing river watershed, and spatial heterogeneity of different landscape metrics was distinct, The following order:LP> HP> MP was observed in the spatial variation of NP, MPS, LPI, DI and CON, and the order for FN, ED, SHDI and SHEI was MP> HP> LP. Without HP area sites, NO3--N had no significant correlation with all landscape metrics; by contrast LPI, DI and CON were extremely significant negatively correlated with the export of NH4+-N and TN, and ED, SHDI and SHEI were extremely significant positive, as a result ED played a crucial factor in spatial variation of NH4+-N and TN; MPS and DI had a significant negative impact on the output of DIC, and for SHDI and SHEI a significant positive impact was found, and then DI played a crucial factor in spatial variation of DIC. Compared with landuse structure, only a few landscape metrics could be used for forecasting the export of NH4+-N, TN and DIC, most relationships between landscape metrics and export of non-point source pollution were feeble than land use structure.