http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Paule-Mercado, Ma. Cristina A.,Salim, Imran,Lee, Bum-Yeon,Memon, Sheeraz,Sajjad, Raja Umer,Sukhbaatar, Chinzorig,Lee, Chang-Hee Elsevier 2018 Ecological Indicators Vol.93 No.-
<P><B>Abstract</B></P> <P>Understanding the influence of land use and land cover (LULC) change in stormwater runoff is important for watershed management. In this study, integration of 31 storm events, monthly monitoring of LULC change, Pearson’s correlation, multiple linear regression analysis (MLR) and Personalized Computer Storm Water Management Model (PCSWMM) were applied to quantify the influence of LULC change on stormwater quality from mixed LULC catchment with ongoing land development in Yongin, South Korea. Due to ongoing land development in the catchment, bare land and urban LULC were exponentially increased while agriculture, forest, grassland and water LULC decreased in spatial extent. The correlation analysis showed that stormwater quality was positively correlated to bare land (0.595; Cl – 0.891; TSS, <I>p</I> < 0.05) and urban (0.768; TN – 0.987; TSS, <I>p</I> < 0.05); negatively correlated to forest (−0.593; Cu – 0.532; BOD<SUB>5</SUB>, <I>p</I> < 0.05) and grassland (−0.587; TSS – 0.512; BOD<SUB>5</SUB>, <I>p</I> < 0.05) and; either positively or no correlation to agriculture (0.064; Cu – 0.871; TSS, <I>p</I> < 0.05) and water (−0.131; Cl – 0.221; TP, <I>p</I> < 0.05). Furthermore, the MLR analysis showed that combinations of different LULC were able to describe the overall stormwater quality of the catchment. Moreover, the LULC scenario analysis demonstrate that under dominant agriculture (S1), bare land (S2) and urban areas (S5), the average pollutant concentrations would increase by as much as 13.22% (Cl; S2; pre-) to 59.25% (TSS; S5; early-active); while under dominant forest (S3) and grassland (S4) the average pollutant concentration would decrease by as much as −53% (Pb; S3; late-active) to −3.22% (BOD<SUB>5</SUB>; S4; pre-). These findings explained that the variability of pollutant concentrations in different phase of land development was affected by expansion of bare land and urban spatial extent, increase of hydrological characteristics (total rainfall and average rainfall intensity) and massive soil activities (soil digging and soil transfer). Therefore, results of this study will provide scientific information to establish a cost-effective stormwater management, development of empirical model, and designing monitoring strategies and guidelines to minimize the negative impact of LULC change on stormwater runoff.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Long-term LULC and stormwater monitoring advanced the current watershed management. </LI> <LI> PCSWMM was used to evaluate the influence of land development on stormwater runoff. </LI> <LI> Land development influences the variability of pollutant concentration in runoff. </LI> <LI> Conversion of vegetation to bare land and urban is the major stormwater stressor. </LI> <LI> Expansion of vegetation cover was not enough to achieve the water quality criteria. </LI> </UL> </P> <P><B>Graphical abstract</B></P> <P>[DISPLAY OMISSION]</P>
Application of neural network for stormwater runoff classification from mix landuse site
( Sheeraz Memon ),( Ma. Cristina Paule ),( Raja Umer Sajjad ),( Imran Saleem ),( Seung-hoon Yu ),( Bum-yeon Lee ),( Chinzorig Sukhbaatar ),( Chang-hee Lee ) 한국물환경학회(구 한국수질보전학회) 2015 한국물환경학회·대한상하수도학회 공동 춘계학술발표회 Vol.2015 No.-
Predictive models play an important part in storm water monitoring due to impact on receiving waters and high cost for data collection. In this study, a neural network approach was used to characterize large amount of stormwater data from an outlet of multiple landuse sites including urban and construction in Korea. Out of total 400 data samples, 85% of the data was used for training the network while remaining 15% was used to test the model which is 1/7th of randomly data samples. An attempt was made using coefficient of determination and average relative error values to develop optimized network model. The results revealed that R2 for both estimation and validation varied significantly on different nodes configurations, whereas least generated average relative error for all output constituents were exhibited in the 8 hidden neurons configuration and therefore it was selected for further classification and comparison. The model performance was compared with multiple linear regression model results using R2 values and Nash coefficient. From the results, it was observed that ANN model produced better results because values of MLR were low for all of the constituents. From the findings, it can be suggested that neural network can be used for stormwater quality data.
Lee, Bum-Yeon,Park, Shin-Jeong,Paule, Ma. Cristina,Jun, Woosong,Lee, Chang-Hee Wiley (John WileySons) 2012 Water environment research Vol.84 No.8
<P>The extent of impervious cover in a watershed has been linked to the quality of an urban aquatic environment. The Kyeongan watershed in South Korea was investigated to evaluate the relationship between the total impervious area (TIA) and the aquatic ecosystem of the watershed, including water quality and aquatic life using a relatively high-resolution (0.4 m) image. The TIA was found to be approximately 12% of the watershed, which indicates that the quality of its environment was being adversely affected by it. For water quality, Pearson correlation analyses showed that all water quality parameters studied were found to be positively correlated with TIA at p < 0.01, except for nitrate (NO3-). In addition, the zone with a higher TIA was found to have worse water quality. Some water quality parameters, such as nitrite (NO2-), total phosphorus, and phosphate (PO4(3-)) were highly affected by discharges from wastewater treatment plants. Water quality data suggest that TIA could be used to predict the water quality of streams. For ecological parameters, the diatom index for organic pollution and trophic diatom index were found to be highly correlated with TIA, whereas physical habitat and benthic macroinvertebrates were poorly correlated with TIA. However, the results indicate that the extent of impervious cover can be a useful indicator for predicting the status of specific ecosystem of streams.</P>