[Influence of Landscape Heterogeneity on Total Nitrogen Concentration in Zhongtian River Watershed].

Nutrient loading into rivers is generally changed by human-induced land use changes and can lead to surface water quality changes. Understanding the extent to which landscape heterogeneity influences nutrient loading is critical to the development of best management practices aimed at water-quality improvement. In this study, the year-round concentrations of total nitrogen were monitored from January 2013 to December 2013 in the 20 hydrological stations of Zhongtian River. Considering the nested relationship of watershed unit, seven land use structures and 13 landscape pattern indexes were chosen, and Pearson correlation analysis, principal component analysis and multiple regression analysis were used to explore the effects of land use pattern on total nitrogen concentrations in Zhongtian River. The results showed that: (1) Construction land and grasslands had a significant impact on total nitrogen concentration, construction land would worsen the water quality, while grasslands could improve water quality; (2) Landscape patch- shape index, landscape patch-area index and construction land were the main factors affecting the total nitrogen concentration in landscape scale; (3) Regression analysis showed that the construction land and landscape patch-area index contributions to total nitrogen concentrations were 67.31% and 32.69%, respectively.

[Sensitivity analysis of AnnAGNPS model's hydrology and water quality parameters based on the perturbation analysis method].

Sensitivity analysis of hydrology and water quality parameters has a great significance for integrated model's construction and application. Based on AnnAGNPS model's mechanism, terrain, hydrology and meteorology, field management, soil and other four major categories of 31 parameters were selected for the sensitivity analysis in Zhongtian river watershed which is a typical small watershed of hilly region in the Taihu Lake, and then used the perturbation method to evaluate the sensitivity of the parameters to the model's simulation results. The results showed that: in the 11 terrain parameters, LS was sensitive to all the model results, RMN, RS and RVC were generally sensitive and less sensitive to the output of sediment but insensitive to the remaining results. For hydrometeorological parameters, CN was more sensitive to runoff and sediment and relatively sensitive for the rest results. In field management, fertilizer and vegetation parameters, CCC, CRM and RR were less sensitive to sediment and particulate pollutants, the six fertilizer parameters (FR, FD, FID, FOD, FIP, FOP) were particularly sensitive for nitrogen and phosphorus nutrients. For soil parameters, K is quite sensitive to all the results except the runoff, the four parameters of the soil's nitrogen and phosphorus ratio (SONR, SINR, SOPR, SIPR) were less sensitive to the corresponding results. The simulation and verification results of runoff in Zhongtian watershed show a good accuracy with the deviation less than 10% during 2005- 2010. Research results have a direct reference value on AnnAGNPS model's parameter selection and calibration adjustment. The runoff simulation results of the study area also proved that the sensitivity analysis was practicable to the parameter's adjustment and showed the adaptability to the hydrology simulation in the Taihu Lake basin's hilly region and provide reference for the model's promotion in China.

[Source analysis of urea-N in Lake Taihu during summer].

To study the effect of urea nitrogen on the ecosystem of Lake Taihu, we conducted urea and various nitrogen analysis for the water samples collected from the lake and surrounding rivers during summer. The ecological index analysis of 82 sites in rivers and lake yielded the following results: (1) The urea nitrogen contents in Taihu ranged from 0.011 to 0.161 mg x L(-1), which was high in the northwest and low in the southeast, related to the main pollution sources distribution of its drainage basin. (2) The dissolved nitrogen was dominated by inorganic nitrogen and the ratio between ammonia nitrogen and nitrate nitrogen was 5: 1. The average percentage of urea nitrogen in total nitrogen, dissolved nitrogen, dissolved organic nitrogen and bioavailable nitrogen was respectively 2.28%, 5.91%, 15.86%, and 6.22%, which showed a significant ecological function in Taihu. (3) Urea nitrogen concentration in river was more than twice that in lake, and the lake river concentration was slightly higher than the river into the lake. (3) In Taihu, there was a transformation relationship between urea nitrogen and the nitrogen in other forms. It showed that urea nitrogen had a significant positive correlation with permanganate index and the other forms of nitrogen, and a significant negative correlation with dissolved oxygen. In addition, urea nitrogen was weakly and positively correlated with chlorophyll a, while closely related to the spatial distribution of benthos and zooplankton species. All the results above showed that urea nitrogen was the bridge of organic and inorganic nitrogen transformation, and was the sign of nitrogen cycle of Lake Taihu, which was controlled by the circulating rate. High nitrogen content (especially the organic nitrogen) and low dissolved oxygen content were the key contributors to the increased urea nitrogen content. In Taihu, the urea nitrogen content was affected by both exogenous input and endogenous release.

[Parameter sensitivity analysis of runoff simulation and model adaptability research based on HSPF].

Based on the sensitivity analysis of key parameters, we calibrated and validated HSPF model in a typical small watershed of hilly region in Tai Lake, and then evaluated the adaptability of the model in the hilly region of Tai Lake. In this study, 5 sensitivity parameters AGWRC, UZSN, INFILT, LZSN, and DEEPFR were determined by using the relative sensitivity method which is based on perturbation analysis, and then the annual runoff from 2005 to 2010 was simulated. The results showed good agreement with percent differences of - 6.39%, - 8.76% and 9.66%, respectively, which were all below 10% during the calibration period, and the simulation results were in a good range with percent differences of -13.31% and 5.22% during the validation period. The Nash-Sutcliffe (Ens) coefficients reached 0.87 and 0.69, respectively, and relative errors were 1.63% and 4.14%, which met the simulation requirements and could reflect the variation of annual runoff in Zhongtian River Basin. The results show that the HSPF model in the study area has better adaptability to watershed runoff simulation which has scientific significance to the simulated hydrological processes in the region and also provides reference for application and promotion of HSPF model in different regions in our country.

[Impact on nitrogen and phosphorous export of wetlands in Tianmu Lake watershed].

Focused on understanding the function of wetland in improving water quality, Pingqiao watershed and Zhongtian watershed in Tianmu Lake drinking water sources area were selected as the research region. We integrated remote sensing, GIS techniques with field investigation and chemical analysis to analyze the relationship between wetland and water quality in watershed scale. Results show: (1) There are many wetland patches in Pingqiao and Zhongtian watershed, wetlands patch densities were respectively 7.5 km(-2) and 7.1 km(-2). Wetlands widely distributed in the Pingqiao watershed with mostly located away from the river of 500 m, whereas wetlands relatively concentrated in the lower reach within 500 meters of riverside in Zhongtian watershed. (2) Nitrogen and phosphorus nutrient retention of wetland in watershed scale was significant. The annual mean TN and DTN concentration had a strong relationship with percent area of wetlands in Zhongtian watershed while the weakest relationship was found with TP and DTP concentrations, especially, the mean TN and DTN concentrations in spring and winter had the significantly negative relationship with wetland areas of watershed. The negative relationship was existed for nitrogen in autumn of Pingqiao watershed, which suggested that watersheds varying in area of wetlands have the different nutrient reducing efficiency in seasonal periods. (3) A certain number and area of wetland will improve river water quality in watershed scale, which can instruct water environment treatment. However, considering the complexity of nutrient transport processes in watershed, wetland-related factors such as area, location, density, ecosystem structure and watershed-related factors such as temporal interval, spatial scales, slope and land use will impact on the transport processes, and related theoretical and practical problems need further research.

[Review on HSPF model for simulation of hydrology and water quality processes].

Hydrological Simulation Program-FORTRAN (HSPF), written in FORTRAN, is one ol the best semi-distributed hydrology and water quality models, which was first developed based on the Stanford Watershed Model. Many studies on HSPF model application were conducted. It can represent the contributions of sediment, nutrients, pesticides, conservatives and fecal coliforms from agricultural areas, continuously simulate water quantity and quality processes, as well as the effects of climate change and land use change on water quantity and quality. HSPF consists of three basic application components: PERLND (Pervious Land Segment) IMPLND (Impervious Land Segment), and RCHRES (free-flowing reach or mixed reservoirs). In general, HSPF has extensive application in the modeling of hydrology or water quality processes and the analysis of climate change and land use change. However, it has limited use in China. The main problems with HSPF include: (1) some algorithms and procedures still need to revise, (2) due to the high standard for input data, the accuracy of the model is limited by spatial and attribute data, (3) the model is only applicable for the simulation of well-mixed rivers, reservoirs and one-dimensional water bodies, it must be integrated with other models to solve more complex problems. At present, studies on HSPF model development are still undergoing, such as revision of model platform, extension of model function, method development for model calibration, and analysis of parameter sensitivity. With the accumulation of basic data and imorovement of data sharing, the HSPF model will be applied more extensively in China.

[Impacts on nutrient export by landscape heterogeneity based on sub-watershed].

Landscape features of a watershed are important factors affecting non-point source (NPS) pollution. Sub-watershed bounds were delineated and landscape heterogeneity was analyzed based on GIS and RS in Xitiaoxi watershed which located the upper reach of Taihu Lake area. Nutrient export intensity of sub-watersheds was estimated by revised export coefficient model. Then the relationships between nutrient export and main landscape types, as well as Shannon diversity index (SHDI) in sub-watershed units were analyzed. Results show, TN and TP export intensity have obvious spatial difference, which changed from 3.01 kg/(hm2 x a) to 15.44 kg/(hm2 x a) and 0.049 kg/(hm2 x a) to 0.355 kg/(hm2 x a) respectively. The dominated landscape types including cultivated land and forest land quantitatively related with nutrient export intensity. TN and TP export intensity will decrease 0.203 1 kg/(hm2 x a) and 0.0152 kg/(hm2 x a) respectively with 10% increased of forest area, and will increase 0.5726 kg/(hm2 x a) and 0.0273 kg/(hm2 x a) with 10% increased of cultivated land area. The relationship between nutrient export intensity and SHDI exhibited second-degree polynomial, export intensity increased by SHDI increasing and to maximum when SHDI equals 1.5, then decreased with SHDI increasing. This research results will provide an important reference value for NPS management.

[Estimated nutrient export loads based on improved export coefficient model in Xitiaoxi watershed].

It is very important to estimate nutrient export loads from watershed, which is beneficial to water environmental management. The export coefficients for major land use types were evaluated by field surveying data, in addition the spatial distributions of precipitation and runoff were considered, to improve on export coefficient model (ECM), then integrated ECM into geographical information system to compose half-distribution model of export coefficients. Estimated nutrients export loads based on the ECM in Xitiaoxi watershed located upper of Taihu Lake area. Results show that the export load for TN and TP was 2 121.3 t and 49.3 t respectively, which provides important reference value for nonpoint source pollution management.