of county-level census and economic data between ... and Morrison; and in Wisconsin they were. Adams, Douglas, Florence, Marathon, Vilas, ...... using GIS.
(Received 26 February 2005; in final form 8 August 2005) .... measurements of the environmental influence of urban land-use zoning. 2. .... 1z lTB/r. Ð Ñlne. Ð3Ñ where l is the wavelength of emitted radiance (for which the peak response and.
Mar 2, 2018 - tionship between land surface IR emissivity and soil moisture. ... and the MetOp Advanced Scatterometer (ASCAT).13 In addition, the recent.
logit and binary logistic models to measure travel mode choice and decision to internalize trips, the evidence here ... INTRODUCTION. Assumptions about travel behavior feed directly into the travel modeling process, ... Newer models integrating trans
Jun 30, 2014 - atmospheric conditions (water vapor and temperature). At present, several ... water vapor content must be known for the exact time of image acquisition, as well as the knowledge of ...... sistema-informacion-agroclimatica-regadio/Red-E
Za dobivanje NDVI, NDBI i NDWI i temperature povrÅ¡ine koriÅ¡tene su satelitske slike. Landsata 5 TM. .... m) and free download availability of website of US.
a clinometer and a compass. The aspect was later converted into soil moisture classes, ranging from 1 (dry) to 20 (wet). These discrete soil moisture classes ...
use of Twitter data in urban studies, especially in relation to land uses. Section 3 reports the ... Section 4 contains an analysis of the results and section 5 presents the main ... 1â2 percent of the whole stream of tweets (Andrienko et al., 2013
Follow this and additional works at: http://digitalscholarship.unlv.edu/sea_fac_articles ... Transactions on Geoscience and Remote Sensing, 48(1), 491-498.
monocarbide. J Mater Sci 1982;17:2657â60.  Kelly A, Groves GW. Crystallography and crystal defects. Lon- don: Longman; 1973.  Stoll WM, Santhanam ...
13909. 15407. Clark. 3544 1323 3075. 3069. 3094. 3428. Custer. 336. 125 291. 291. 293 ... Total nitrogen concentrations in septic tank effluent range from 25 ...
and the R.M. Young Wind Sentry Anemometer (both. Campbell Scientific, Inc. ..... that when calculating these metrics for wind direction, the WRF-simulated or ...
transmission oil cooler honda insight manual transmission for sale honda jazz 2013 manual transmission 1999 honda civic manual transmission noise honda prelude manual transmission fluid change honda accord with manual transmission for sale honda 3.5
The concept of âcognitive fusionâ refers to the extent to which a person believes his thoughts and ... Relationship between drug use and psychopathological variables of risk in university students. 435 al., 1986). This instrument ..... y no clÃn
likelihood algorithm implemented in Idrisi Kilimanjaro software. Accuracy of classified maps was evaluated using 150 sample points sistematically distributed.
The popularity of Facebook and other online social-networking sites has led to research on the potential risks of use, including Internet addiction. Previous ...
May 1, 2018 - LULC Type. Area in ha (Percentage Change). 2015. 2030. Baseline. BAU .... We modeled the LULC changes using statistical projection and ...
understanding about processes and seek to build models that simulate outcomes in specific places. .... engaged in agricultural activity (e.g., Lynam 2003) or residential development (e.g.,. Bockstael 1996). .... As with all fitting techniques, the sp
Aug 25, 2014 - Predicted tumor doses for 50% control (TD50) in 2 Gy fractions were estimated for each assumed uptake profile and for various possible cell ...
Jul 29, 2008 - 1986), traps coarse particulate organic matter and sediment ... fastest growing county in the state; some of the most rapid housing ... Both geography and land use divide the watershed into two parts. .... each land class to characteri
Amol D. Vibhute, Dr. Bharti W. Gawali / International Journal of Engineering Research and. Applications .... types of systems: elements, attributes, and ...... Using IKONOS Images, Agricultural and. Biosystems Engineering Department,. University of P
Nov 1, 2011 - selection, since some studies take into account only those .... Hamilton and Hamilton (1997) found empirical evidence about the hypothesis, stating ..... includes all the individual level answers on vocal questions. ...... available at
Jul 8, 2010 - and the relationship among media use, BMI, socio-demographic profiles, and snacking ..... great deal of time indoors rather than outdoors,.
and Decarrico (1992) and Thornbury (2002) have underlined the importance and benefits of collocations. They have also mentioned that collocations are essential for EFL learners to boost their communicative competence, enhance their fluency, be native
Journal of Environmental Management (2002) 66, 377±393 doi:10.1006/jema.2002.0593, available online at http://www.idealibrary.com on
Modeling the relationship between land use and surface water quality Susanna T. Y. Tong* and Wenli Chen Geography Department, University of Cincinnati, Cincinnati, OH 45221-0131, USA Received 21 September 2001; accepted 29 April 2002
It is widely known that watershed hydrology is dependent on many factors, including land use, climate, and soil conditions. But the relative impacts of different types of land use on the surface water are yet to be ascertained and quanti®ed. This research attempted to use a comprehensive approach to examine the hydrologic effects of land use at both a regional and a local scale. Statistical and spatial analyses were employed to examine the statistical and spatial relationships of land use and the ¯ow and water quality in receiving waters on a regional scale in the State of Ohio. Besides, a widely accepted watershed-based water quality assessment tool, the Better Assessment Science Integrating Point and Nonpoint Sources (BASINS), was adopted to model the plausible effects of land use on water quality in a local watershed in the East Fork Little Miami River Basin. The results from the statistical analyses revealed that there was a signi®cant relationship between land use and in-stream water quality, especially for nitrogen, phosphorus and Fecal coliform. The geographic information systems (GIS) spatial analyses identi®ed the watersheds that have high levels of contaminants and percentages of agricultural and urban lands. Furthermore, the hydrologic and water quality modeling showed that agricultural and impervious urban lands produced a much higher level of nitrogen and phosphorus than other land surfaces. From this research, it seems that the approach adopted in this study is comprehensive, covering both the regional and local scales. It also reveals that BASINS is a very useful and reliable tool, capable of characterizing the ¯ow and water quality conditions for the study area under different watershed scales. With little modi®cation, these models should be able to adapt to other watersheds or to simulate other contaminants. They also can be used to study the plausible impacts of global environmental change. In addition, the information on the hydrologic effects of land use is very useful. It can provide guidelines not only for resource managers in restoring our aquatic ecosystems, but also for local planners in devising viable and ecologically-sound watershed development plans, as well as for policy makers in evaluating alternate land management decisions. # 2002 Elsevier Science Ltd. All rights reserved.
Keywords: BASINS, ArcView, GIS, watershed hydrologic modeling, water quality, ¯ow.
Introduction As water drains from the land surface, it carries the residues from the land. Surface runoff, especially under the ®rst ¯ush phenomena, is an important source of non-point source pollution. Runoff from different types of land use may be enriched with different kinds of contaminants. For example, runoff from agricultural lands may be enriched with nutrients and sediments. Likewise, runoff from highly developed urban areas may be enriched with rubber fragments, heavy metals, as well as sodium and sulfate from road deicers. Moreover, * Corresponding author. E-mail: [email protected] 0301±4797/02/$ ± see front matter
through evapotranspiration, interception, in®ltration, percolation and absorption, different types and coverages of vegetative surfaces can modify the land surface characteristics, water balance, hydrologic cycle, and the surface water temperature (LeBlanc et al., 1997). As a result, the quantity of water available for runoff, stream¯ow and ground water ¯ow, as well as the physical, chemical and biological processes in the receiving water bodies can be affected. It is therefore conceivable that there is a strong relationship between land-use types and the quantity and quality of water (Gburek and Folmar, 1999). In a study of the effects of forested, agricultural and urban areas on water quality and aquatic biota in the Piedmont ecoregion of North Carolina, Lenat # 2002 Elsevier Science Ltd. All rights reserved.
S. T. Y. Tong and W. Chen
and Crawford (1994) found that the agricultural lands produced the highest nutrient concentrations. Fisher et al. (2000) also noted a higher amount of nitrogen, phosphorus and Fecal coliform bacteria in the poultry production areas in the Upper Oconee Watershed in Georgia. In another study of Coweeta Creek in western North Carolina, Bolstad and Swank (1997) observed that there were consistent changes in water quality variables, concomitant with land-use change. Similarly, in an earlier study of the Little Miami River Basin, Tong (1990) found that urban development in the watershed had caused substantial modi®cation on ¯ood runoff and water quality. Changing land use and land management practices are therefore regarded as one of the main factors in altering the hydrological system, causing changes in runoff (Mander et al., 1998), surface water supply yields (Wu and Haith, 1993), as well as the quality of receiving water (Changnon and Demissie, 1996). Although there have been some studies on the impacts of land use on water ¯ows and quality (see for example, the work of Hanratty and Stefan, 1998; Rai and Sharma, 1998; and Bhaduri et al., 2001), the complex intrinsic relationships of land use, water yields and water quality in different geographical areas under different scales are yet to be elucidated. Current methods on predicting water quality in river catchments based on land-use patterns are still developmental. Most existing research is con®ned to ®eld studies. Some studies are very speci®c to a locality at one geographical scale. Many focused on either statistical, or spatial, or modeling analyses. Others examined the impacts of land use on only the quantity or the quality aspect of runoff. Examples of such work include those performed by Meissner et al. (1999), Ferrier et al. (1995), Tsihrintzis and Hamid (1998), Mattikalli and Richards (1996), Wu and Haith (1993), Hulme et al. (1993), Henderson-Sellers (1994), and Bouraoui et al. (1998). Only a few studies have employed an integrated approach involving the use of statistical and spatial analyses, as well as hydrologic modeling to examine the hydrologic effects of land use in both a regional and a local scale.
Study objective The objective of this study was to use a watershedbased approach to examine the plausible statistical and spatial relationships of land use on the quantity and quality of the surface water under a broad regional scale in the State of Ohio, and to model the
relative impacts of different types of land use in a local watershed. The aim is to further our understanding of the hydrologic effects of land use.
Methods and materials Both statistical (such as non-parametric correlation analysis and analysis of variance) and geographic information system (GIS) analyses were adopted in this study. They were used to examine the general association of land use and ¯ow and water quality, and locate the watersheds that are enriched with contaminants and with strong relationships with land use. GIS is a powerful data integration and spatial analysis tool. In this research, ArcView GIS is used to aggregate, synthesize and analyze large databases, and to identify spatial relationships. Based on the correlation results and the availability of data, a watershed was chosen for detailed modeling of the relative hydrologic impacts of different types of land use in a local watershed scale. A sophisticated and widely used assessment tool, the Better Assessment Science Integrating Point and Nonpoint Sources, BASINS, (US Environmental Protection Agency, 1998) was utilized in this study. BASINS is a physical-process-based analytical system developed by the US Environmental Protection Agency (USEPA) as an assessment tool for watershed-based and water-quality-based studies. Version of BASINS 201 was used in this paper to characterize the ¯ow and water quality conditions in the watershed. Several water quality parameters (nitrates, phosphates, and Fecal coliform) were examined in detail. The models were calibrated and validated using ¯ow and water quality data from historical records retrieved from US Geological Survey (USGS). The ef®cacy of the models in representing the real world runoff and water quality conditions under different geographical scales was also assessed. The approach to the analyses entailed ®rst of all the establishment of whether there are any statistical and spatial relationships between land use and water quality at a regional scale. If such relationships exist, the second task was to model the water quality from a local watershed with different types of land use. The State of Ohio was selected in this project for the broad scale regional study due to the availability of historical data on ¯ow, water quality, land use, elevation, and climate. In the hydrololgic modeling, a watershed from the Little Miami River Basin, the East Fork Little Miami River Basin, was chosen.
Land use and surface water quality
Statistical and spatial analysis Statistical analyses were utilized to test the null hypothesis that water quality is not related to the surrounding land-use types at a regional scale. They were also used to determine the quantitative associations, if any, of land use and water quality in the State of Ohio. The GIS spatial analyses were employed to map the areas that have a high contamination level and to examine the spatial relationships of land use and water quality.
Data preparation The map of the 11-digit Hydrological Units (HUCs) for the State of Ohio was obtained from USGS. It was used as the base map for the analyses. The water quality data for the period of 1990±1998 in Ohio were obtained from the USEPA's Storage and Retrieval computerized environmental data system (STORET). STORET is a repository for water quality, biological, and physical data. The data are in database (DBF) format. There are 243 water quality variables. Each monitoring site is referenced in latitude and longitude. To spatially locate these monitoring stations, an Arc Avenue program was developed to extract the water quality information using the longitude and latitude values of each monitoring station. Using the spatial location function of ArcView, a shape ®le was created for all the monitoring stations. The original water quality table was then joined to this spatial map by station; as such queries on the data could be performed and the water quality data could be geo-referenced spatially by each station. The map was projected into Albers map projection to match the 11-digit HUCs base map. By employing the spatial selection in ArcView, all the stations in each HUC were obtained. The mean values of each water quality variable in each 11-digit HUC were then calculated. These constituted the water quality data for statistical and spatial analyses. Land-use data for the State of Ohio in 1994 were obtained from Landsat Thematic Mapper dataset acquired by the Multi-Resolution Land Characterization Consortium (MRLC). The MRLC data are in a grid format, with land-use types listed for each pixel. Since the HUCs are displayed in a vector format, ®le conversion was performed. The `add' function in ArcView GIS was used to combine the two ®les. The percentages of each type of land use in each HUC were then calculated.
Prior to the statistical analyses, the normality and linearity of the water quality and land-use data were tested. Only pH and sulfate water quality variables and forest land-use variable exhibited some form of normal distribution. Scattergrams also revealed that the relationships of land use and water quality might not be linear. This latter result conforms with other researchers, such as Nikolaidis et al. (1998), who suggested a non-linear response of nutrient loadings to different land-use types due to the nature of hydrogeochemical processes in a watershed. Moreover, in the data sets for total phosphorus, total nitrogen, lead, manganese and Fecal coliform, there are a few outliers (in the form of extremely high values). These outliers as well as missing data and `zero' values were deleted from the data set for further analyses. Only those HUCs containing water quality data were used in the statistical analyses. The analyses Since most of the variables did not distribute normally, the statistical analyses were con®ned to non-parametric statistical tests. Spearman's rank correlation analyses were used to explore the relationships between land use and water quality in the State of Ohio. Analysis of variance on the ranked values of a few selected water quality variables was also performed to test if there were any signi®cant differences between land-use categories. Water quality variables that were signi®cantly related to land use were chosen for further GIS spatial analyses. Based on the frequency distribution, and the mean and median values of each water quality variable, the HUCs were classi®ed into three arbitrary categories, the `High', `Medium', or `Low', depicting the HUCs with a high, medium, or low value of that water quality variable. Similarly, for the land-use variables, the HUCs were classi®ed into `High', `Medium', or `Low', representing the HUCs with a high, medium, or low percentage of certain land-use type. According to the rank correlation results, pairs of water quality and land-use variables that were signi®cantly related were selected. Each pair of water quality and land-use variables was superimposed using ArcView GIS. For example, if nitrogen was positively related to agriculture land use, the HUCs with high nitrogen values were overlaid with the HUCs with a high percentage of agriculture land use. The resultant map denoted the spatial relationships of that water quality variable with that land-use type.
S. T. Y. Tong and W. Chen
Hydrologic modeling on the East Fork Little Miami River watershed The selection of a hydrologic model There are different types of hydrologic models in use today. They range in capability, complexity, scale and resolution. With various theoretical assumptions and mathematical algorithms, these models may have different data requirements. Moreover, the accuracies of the resulting simulation may vary (Singh, 1995). The most common full-scale simulation models for urban areas are L-THIA (Bhadhuri et al., 1997), Sacramento Soil Moisture Accounting Model (Burnash et al., 1973), SWAT (Arnold et al., 1994), SWIM (Krysanova and Luik, 1989), HYDROTREND (Syvitski et al., 1988), System Hydrologique Europeen Model (Abbott et al., 1986), and the Hydrological Simulation ProgramFORTRAN (Johanson et al., 1984; Donigian et al., 1984). Other models for non-point source pollution include CREAMS (Knisel, 1980), GLEAMS (Leonard et al., 1987), SWRRB (Arnold et al., l990), MATSALU (Krysanova et al., 1998), ANSWERS (Beasley et al., 1980), and AGNPS (Young et al., 1989). A number of hydrologic models were considered in this research. The main criteria for choosing the model were: model accuracy, model capabilities, model ¯exibility, data requirements, and ease of use. Based on extensive literature searches for models in use today and communications with other model users, BASINS was chosen to model the quantity and quality of the runoff from different types of land use.
BASINS BASINS is a ¯exible, yet valuable watershed-based multipurpose integrated water-quality analysis system. Using the familiar Windows environment and an ArcView-based GIS as an integrating framework, BASINS incorporates commonly used environmental data, a variety of proven and robust analytical hydrologic, point and non-point loading, and water quality stream models, assessment and planning tools, statistical analytical tools, data mining, organizing, evaluating and management technologies, as well as reports and graphics display and visualizing capabilities into one system. Accordingly, it can support analysis at a variety of scales and can integrate and display a wide range of information in various formats.
There are two stream models (QUAL2E and TOXIROUTE) and a Nonpoint Source Model (NPSM) in BASINS. The NPSM is a planninglevel watershed model. By incorporating both point and non-point sources, NPSM is capable of simulating non-point source runoff and associated pollutant loadings, point source discharges, ¯ow and water quality routing through stream reaches and well-mixed reservoirs. NPSM uses most of the simulation capabilities of the Hydrological Simulation ProgramÐFORTRAN (HSPF) hydrologic model (US Environmental Protection Agency, 1984). HSPF is a very robust, high resolution, ¯exible, reliable, and comprehensive hydrologic model for simulation of watershed hydrology and water quality (Bicknell et al., 1996). As a physical-process-based model, HSPF uses minimal input data to describe hydrological conditions in a watershed. As a time series management system, it can simulate continuously the hydrologic and associated water quality processes on pervious and impervious land surfaces as well as in streams and well-mixed impoundments. Derived from the Stanford Watershed Model (SWM), HSPF considers all stream¯ow components (runoff, inter¯ow, and base¯ow) and their pollutant contributions. In addition, it has incorporated many non-point source models, such as ARM and NPS. By integrating the chemical, biological, and contaminant runoff processes on land surfaces and in the soil pro®les with in-stream hydraulic, water temperature, sediment transport, and nutrient and sediment-chemical interactions, it simulates hydrolysis,oxidation,photolysis, biodegradation, volatilization and sorption. Based on a continuous record of precipitation and evaporation data, it computes a continuous hydrograph of stream¯ow at the basin outlet and produces a time history of the runoff, sediment load, and nutrient and pesticide concentrations (Donigian and Huber, 1991). Indeed, it is an integrated and comprehensive model for runoff, erosion, fate and transport of pollutants. It has been widely-used for simulating watershed hydrology and water quality, and has been applied to support various watershed and water quality modeling studies. HSPF is a continuous simulation program. It requires continuous records of rainfall, evapotranspiration, temperature, and solar intensity to drive the simulations. Thus, only those watersheds which are close to a meteorological station with reliable long-term climate data can be used in the study. Moreover, for calibration purposes, the watershed has to have USGS gauge stations that have historical discharge, ¯ow and water quality information.
Land use and surface water quality
The East Fork Little Miami watershed In this project, the East Fork Little Miami River, a tributary of the Little Miami River in southwestern Ohio, was chosen for further hydrological modeling. This was mainly because the Little Miami River has been monitored by the USGS since 1920. Continuous records of stream ¯ow are available for many areas in the river. In addition, there are historical climate, land use, ¯ow, permitted point source discharge and water quality data. The land-use types in the river basin are diversi®ed. Besides, the preliminary statistical and GIS analyses found that there were strong relationships between land use and water quality in this river basin. After calibration and validation, the models were used to characterize the ¯ow and water quality conditions in the watershed. The Little Miami River is a major tributary of the Ohio River. It ¯ows from northeast to southwest and joins the Ohio River near Cincinnati, Ohio. Its watershed covers more than 5840 sq. km and 3970 stream km. The dominant parent materials of the region consist of shale, limestone and dolomite from the Upper Ordovician or Silurian (Ohio Department of Natural Resources, 1964). Most of the soils belong to Genesse-Williamsburg Association and have their origins from alluvial, glacial outwash, residual and glacial till plains. They are deep, moderately well drained and highly productive (Lerch et al., 1975). The sand and gravel deposits provide aquifers with fair yield potentials for groundwater supplies. Containing some of Ohio's most scenic and diverse riverine habitats, the Little Miami River was designated a state scenic river in 1969 and a national scenic river in 1973. In addition, the river is one of the largest `exceptional warmwater habitat stream' in Ohio, supporting a very rich aquatic community of ¯ora and fauna (Debrewer et al., 2000). Notwithstanding the fact that most lands in the river basin are agricultural, there has been a steady increase of residential and commercial land use due to recent suburban sprawls. Natural forests and agricultural lands are gradually replaced by developed areas. The river system is now under stress. Any change in stream ¯ow caused by landuse modi®cation will undoubtedly aggravate the situation, reducing the water availability for municipalities, degrading water quality, and increasing the potentially toxic algae (Ohio Environmental Protection Agency, 1996). In order to conserve the area, it is of paramount importance to have a better understanding of the hydrologic consequences of land use. This type of knowledge will be invaluable
not only to academia, but also to the resource managers and local government. The East Fork Little Miami River is the eastern branch of the Little Miami River. The watershed used in this simulation study is the southern most HUC in the Little Miami River watershed. There are eleven reaches in this HUC. In this East Fork Little Miami watershed, agriculture is the predominant land use, although there are more forest and urban areas in the western part. Database construction and model preparation Since BASINS is a watershed-based model, the watershed for the analyses had to be delineated ®rst. This was done by using the delineation tool in BASINS. The watershed for the East Fork Little Miami Rivers was delineated according to the Reach Files Version 1 and Version 3 (RF1 and RF3, respectively) from the USEPA and digital elevation data (DEM) from USGS. To examine the ef®cacy of BASINS in simulating watersheds at different scale, the East Fork Little Miami River Basin was further partitioned into eleven smaller subwatersheds according to the eleven reaches in the Reach File Version 1 (Figure 1). Each of these smaller subwatersheds is identi®ed using its reach identi®cation code. The use of smaller subwatersheds is helpful because there are subwatersheds in the study area that are dominated by different types of land use, which might impact the hydrology differently. By comparing the output of ¯ow, nitrogen, phosphorus, and Fecal coliform from different land-use types in different subwatersheds, the relative impacts of land use on water quantity and quality could be ascertained. Once the subwatersheds were delineated in BASINS, then GIS data were prepared, the NPSM ®les were created, and a NPSM model speci®c for the study area was built. The GIS data were extracted from the BASINS database using the `Data Extraction' tool in BASINS. These data included connection and topological relationships, water quality observation data, permit point source pollution data, and information on river ¯ow and HUCs. Furthermore, the 1970±1990 land-use coverages were obtained from the MRLC. Meteorological data were obtained from the National Climatic Data Center. Soil data were acquired from the National State Soil Geographic (STATSGO) database produced by the U.S. Department of AgricultureÐNational Resources Conservation Service. All of these additional data were compiled
S. T. Y. Tong and W. Chen
Delineated East Fork Little Miami watershed.
and managed in a GIS database. Spatial coverages were merged and clipped using Arc/Info. They were then imported into the NPSM project. Next, the *.uci ®le (user input ®le) was generated and the most appropriate meteorological station and the Watershed Data Management (WDM) ®le selected. The NPSM was run for the subwatersheds at the lowest pour point.
A basic hydrologic model for the East Fork Little Miami River Basin was then built by ®rst running the HSPF module in NPSM for each smaller subwatershed using the default parameters. According to the connection relation among the reaches, the simulated ¯ow from the subwatershed at the upper reaches was added to the HUC downstream as a point source ¯ow. After simulating all the subwatersheds in the basin, a hydrologic model for the East Fork Little Miami River Basin was created.
Hydrologic simulation The simulation was performed in two steps: the simulation of the hydrology and the simulation of the water quality parameters. This is because water quality simulation is based on the general hydrologic model. Simulation was based on January 1, 1988 to December 31, 1994 meteorological data from the Covington WSO Airport Weather Station in Northern Kentucky, and 1990s land-use data from MRLC. By default, BASINS assumes all built-up areas to have a 50% pervious and a 50% impervious land surface. To re¯ect the actual percentages of pervious and impervious areas in the urban areas, a land-use report for the study area was generated from BASINS. The information was then used to adjust the % of perviousness and % of imperviousness in the watershed.
Model calibrations The simulated ¯ow results for the East Fork Little Miami River, in terms of total rate of out¯ow of reaches, were then compared with the observed daily discharge records during the simulation period from existing USGS monitoring data. It was found that by using the default parameters, the simulated ¯ow was very close to the actual monitored ¯ow rate, with an error rate of about 812% reduction, meaning that the simulated ¯ow was 812% lower than the actual observed USGS ¯ow data (Figure 2). The error rate is calculated as: (Simulated valueÿMonitored value) 7 Simulated value. Generally, if the error rate is less than 10%, the model is regarded to be good. The model is still
Land use and surface water quality
Simulated ¯ow with default input parameters compared with observed ¯ow, East Fork Little Miami River.
acceptable if the rate ranges from 10±20% (Bicknell et al., 1996). In order to better re¯ect the real world conditions, the HSPF model parameters were calibrated until there was a satisfactory agreement between the simulated ¯ow rate and the observed monitored ¯ow data. There are different methods for calibration. BASINS provides some parameter estimates and the users can adjust the input parameters manually. In addition, computer programs like Annie, Parameter Estimation (PEST 2000), and the Expert System for the Calibration of the Hydrological Simulation ProgramÐFortran (HSPEXP) can be used to help estimating the parameters. Since the error rate was so low, calibration was done manually using the HSPF Parameter Database (HSPFParm). The `input data editor' in BASINS was used to modify the input parameters. In the original hydrologic model, the base ¯ow was too low. Thus, the parameters in the `pervious land hydrology' algorithm were modi®ed. According to BASINS Technical Notes (US Environmental Protection Agency, 2000), the `fraction of remaining evapotranspiration for base ¯ow' (BASETP) was reduced. This decreased the evapotranspiration by riparian vegetation. The default setting for BASETP was
originally 002. After several iterations, it was ®nally adjusted to 0001. Under this parameter setting, the error rate for stream¯ow was reduced to 01% reduction. It implies that the simulated results for the watershed are almost the same as the actual monitored values and the hydrologic model for the East Fork Little Miami River Basin can realistically simulate the actual ¯ow conditions. Simulation under a subwatershed scale The simulated ¯ow conditions for ®ve smaller subwatersheds in the river basin were further evaluated. These subwatersheds were selected because each of them is located very near to a monitoring gauge station. The simulated ¯ow results from these subwatersheds were compared with the observed ¯ow records. The results indicated that the ¯ow values were very close to the monitored values and all of them were within acceptable limits. Moreover, when the ¯ow values for each reach were examined, it was observed that the combined ¯ow values in the upper reaches were found to be approximately the same as the ¯ow rate in the lower reach. For example, the simulated ¯ow rate for Reach
S. T. Y. Tong and W. Chen
05090202002 on January 1,1993 was 1093 cu m/s. It was approximately equal to the sum of the ¯ow in the two upper reaches, Reach 05090202003 (822 cu m/s) and Reach 05090202010 (237 cu m/s). These results provided insights into the ef®cacy of the models in representing the real world conditions under different geographic scales. Apparently, the model can effectively simulate the ¯ow conditions in the smaller subwatersheds as well as the larger watershed. Consequently, this hydrologic model was accepted and was used for further water quality simulation.
in NPSM. The new information was imported manually to the project. Then, each model was run for each smaller subwatershed. As in the hydrologic modeling, the simulated results for each water quality parameter were compared to the monitored values from USGS and calibration and validation were performed as described earlier.
Water quality simulation
Results from the statistical analyses indicated that land-use types were signi®cantly correlated to many water quality variables in the watersheds in the State of Ohio at a probability level of 500001 (Table 1). For example, total nitrogen, total phosphorus, and Fecal coliform had strong positive relationships with commercial, residential, and agricultural lands. They were also negatively related to forest land use. In addition, agriculture land use was strongly related to conductivity and pH. It was negatively related to zinc, cadmium, lead, and manganese. Residential and commercial lands were related to sodium, cadmium, lead, conductivity, BOD, and zinc. The analysis of variance on the ranked values of nitrogen and phosphorus showed that both variables exhibited signi®cant differences (with P 00001 level) between the land-use classes (Table 2). The mean ranked in-stream total nitrogen and phosphorus values were much higher in the agricultural watersheds (mean of nitrogen 407, mean of phosphorus 426) than the urban
In simulating the water quality conditions, modeling was con®ned to total nitrogen, total phosphorus and Fecal coliform. This is mainly because the earlier correlation analysis showed that these three variables had strong correlations with landuse types. Besides, there is a lack of available historical monitoring date on other water quality parameters for the watershed and available parameter estimates for calibration purposes. Three water quality models (nitrogen, phosphorus, and Fecal coliform) were built through the pollution related routines in NPSM. In each of the models, the information on the point source discharges from the BASINS database was compared with existing Permit Compliance System metadata set from the Of®ce of Water of the USEPA. There were four point sources of nitrogen and phosphorus that were not included in the BASINS database. Consequently, the BASINS data were edited by using the `point source editor'
Results and discussion Statistical and GIS analyses
Table 1. Results of the Spearman's rank correlation analysis on water quality variables and land-use types in all subwatersheds in the State of Ohioa Water quality variables
Conductivity BOD pH Total nitrogen Total phosphorus Sodium Cadmium Lead Manganese Zinc Fecal coliform a
ones (mean of nitrogen 372, mean of phosphorus 388). Forested areas had a much lower nitrogen and phosphorus rankings (mean of nitrogen 239, mean of phosphorus 252). GIS analyses identi®ed a few watersheds that were contaminated. Watersheds in the Lower Great Miami River, Lower Little Miami River, and the Upper Middle Ohio Laughery River Basins had high levels of nitrogen and phosphorus. Watersheds around Cincinnati, Cleveland, and Akron had high levels of nitrogen, whereas watersheds around Cincinnati and Dayton had high levels of phosphorus, and watersheds around Columbus, Akron, Toledo and Cleveland had high Fecal coliform bacteria (Figures 3±5). These are the watersheds that should warrant our restoration efforts. Such information would be useful to land-use planners, Table 2. Results of the analysis of variance on the ranked values of total nitrogen and phosphorus among different land-use classes in all subwatersheds in the State of Ohio Water quality variables Nitrogen Phosphorus
resource managers and regulatory agencies for assessing the current conditions and prioritizing restoration efforts. Figures 6±9 depict some examples of the spatial relationships of land use and water quality in Lower Great Miami River, Lower Little Miami River, and the Upper Middle Ohio Laughery River Basins by overlaying the maps of high % of agricultural and urban land use with nitrogen, phosphorus and Fecal coliform. These maps showed that for agricultural lands, the spatial relationship with total phosphorus was very prominent. The watersheds just northwest of West Carollton City and northeast of Lebanon had both high % of agricultural land and nitrogen and phosphorus loadings (Figures 6 and 7). Regarding the urban lands, phosphorus again exhibited a more apparent spatial relationship. Watersheds in the northern and eastern Upper Middle Ohio Laughery River Basin and the western Lower Little Miami River Basin around Lebanon and Mason had both a high % of urban land use and high values for phosphorus and nitrogen (Figures 8 and 9). This result is very useful as it reveals that the relationships of phosphorus with agriculture and urban land use are very strong. This association is often neglected, as most of our current restoration and
HUCs with high in-stream nitrogen values in the State of Ohio.
S. T. Y. Tong and W. Chen
HUCs with high in-stream phosphorus values in the State of Ohio.
HUCs with high Fecal coliform counts in the State of Ohio.
Land use and surface water quality
HUCs with high in-stream nitrogen values and high % of agricultural land use in three watersheds in Ohio.
HUCs with high in-stream phosphorus values and high % of agricultural land use in three watersheds in Ohio.
S. T. Y. Tong and W. Chen
HUCs with high in-stream nitrogen values and high % of urban land use in three watersheds in Ohio.
conservation efforts are directed toward the reduction of nitrogen levels, especially from agricultural lands. To protect our watersheds, considerations should be made to lower not only the amount of nitrogen, but also phosphorus, in both the agricultural and urban areas. An example is by implementing nitrogen and phosphorus removal and treatment facilities in our sewage treatment plants.
Hydrologic and water quality modeling The outputs for the simulation were plotted as graphs. An examination of the graphs for the eleven subwatersheds revealed that they portrayed very similar hydrologic patterns. The simulation results also disclosed that the impervious lands produced more than 55 times as much runoff as the pervious lands. For example, in subwatershed 05090202002, the arithmetic mean of surface ¯ow generated from the impervious urban lands was 2659 cm per day. The surface runoff from agricultural lands was only 046 cm per day. Moreover, the surface ¯ow and pollutant graphs for the impervious lands for all
subwatersheds approximated that of the precipitation curves. This may be due to the lack of in®ltration and water storage capacities of the impervious land surface. Any rainfall will be readily converted to storm-water discharge. The increased amount of surface ¯ow also will wash away more contaminants from the land surface. Even small rain is capable of washing the pollutants from the impervious surfaces into the receiving waters. Stream ¯ow and, to a certain extent, water quality is therefore primarily determined by rainfall. Figure 10 shows an example of the nitrogen loading generated from impervious urban lands from subwatershed 05090202002. To examine the relative effects of different types of land use on water quality, the output curves from different types of land use were superimposed. Figure 11 is an example of such curves. This ®gure compares the loadings of nitrogen between the agricultural and pervious urban land use. The relative contributions of pollutants from different types of land use were also expressed in terms of the amount of pollutant per unit area of each type of land use. Results showed that agricultural land use produced the highest and barren land use the least amount of contaminants.
Land use and surface water quality
HUCs with high in-stream phosphorus values and high % of urban land use in three watersheds in Ohio.
Figure 10. Simulated nitrogen loading from impervious urban land use in subwatershed 05090202002.
S. T. Y. Tong and W. Chen
Figure 11. Comparison of the simulated nitrogen loading from agricultural land use and pervious urban land use in subwatershed 05090202002.
Most nitrogen loading was produced from agricultural lands. This ®nding can be illustrated by examining the simulation from June 1, 1992 to June 30, 1992 in subwatershed 05090202008. In this agricultural watershed, each acre of agricultural land produced 25 to 45 kg of nitrogen each day. In general, the amount of nitrogen produced from agricultural lands was about seven times than those produced from impervious urban. It was more than nine times as much as pervious urban land. Pervious urban land use yielded a slightly higher level of nitrogen, about 12 times, than the forest lands. But, it produced 24 times that of the barren lands. In terms of the total amount of nitrogen produced per unit area, the order for the different types of land use was: agriculture 4 impervious urban 4 pervious urban 4 forest 4 barren. However, an examination of the simulation curves during storms revealed a slightly different pattern. The order for nitrogen output in different land-use types during storms was: agriculture 4 impervious urban 4 forest 4 pervious urban 4 barren. That means, agricultural, impervious urban and forested areas produce more nitrogen in times of storms and heavy rainfall.
The simulated model results for phosphorus were very similar to nitrogen in terms of the overall output pattern. Nevertheless, the amount of phosphorus output was much smaller than nitrogen. It was about 10 times less than nitrogen. As an example, in the small subwatershed 05090202002, from July 1, 1992 to July 31, 1992, the nitrogen output for this subwatershed was about 216 kg/sq km-day, whereas the phosphorus output was 20 kg/sq km-day. The order for the amount of phosphorus loadings generated from different landuse types was: agriculture 4 impervious urban 4 pervious urban 4 forest 4 barren. Agricultural land use commonly produced more than six times as much phosphorus than impervious urban, 10 times more than pervious urban, 20 times more than forest, and 154 times that of barren land use. The Fecal coliform counts produced from agriculture land use were about ®ve times greater than that from pervious urban areas, seven times than that of forest, and more than 16 times that of impervious land and 46 times that of barren land. The order for the amount of Fecal coliform counts for different land-use types was: agriculture 4 pervious urban 4 forest 4 impervious urban 4 barren.
Land use and surface water quality
There was also a tendency for Fecal coliform counts to increase in winter and spring seasons and decrease during the summer and fall.
Conclusions In this paper, the analysis involved statistical, GIS, as well as hydrologic modeling. The results showed that, unequivocally, land use was related to many water quality parameters. This relationship was evident in a regional scale (the State of Ohio), both statistically and spatially. In smaller watersheds, the impacts of land use on water quality could be modeled effectively using BASINS. From the model results, it was apparent that runoff from agricultural as well as impervious urban land use had much more nitrogen and phosphorus. This was the case especially after rain storms. These results are in accordance with an earlier ®nding where the subwatersheds of Cincinnati and Columbus in the Ohio River Basin were modeled using L-THIA hydrologic model (Liu et al., 2000). If land use changes in the future, the levels of contaminants will be changed accordingly. Hence, future land development and management should be considered with care. This is especially the case if the land is going to be changed to agriculture or impervious urban lands. With a better land-use planning, we may be able to curtail some of the water quality problems. Using an analytical tool such as BASINS would help us to predict the plausible hydrological consequences of such changes in land use. BASINS has many advantages. It is a very comprehensive water quality analysis tool. As demonstrated from this study, it is capable of modeling the water quantity and water quality conditions under different geographic scales with an acceptable degree of accuracy. All of the simulated values were very close to the actual monitored values. It seems likely that with little calibration and validation, the model can be used to characterize the current discharge and water quality conditions in another watershed under a different geographical scale, in a different region with various landscapes, soils, and vegetative environments. The model also can be employed to predict the future hydrological conditions in times of global environmental change, such as land-use change, climate change, or changes in total mass discharge loads (TMDLs), point source discharge or permit discharge conditions. Other contaminants can also be simulated easily.
The present study integrates different techniques to investigate the impacts of land use, as such, the models and the methodology can provide a simple, but effective, management tool for policy makers. In addition, the information derived from this study can have direct application values to state and local agencies, city planners and government resource managers for de®ning the impacts of land use on water resources and for implementing long-term water planning and management schemes. The research ®ndings can also contribute to the existing knowledge of the plausible hydrologic implications of land use.
Acknowledgment Most of the work reported in this paper was funded by the U.S. Environmental Protection Agency under Cooperative Agreement No. CR827679-01-0. The authors are thankful to Dr. Susan Cormier of the National Exposure Research Laboratory (NERL) of the USEPA in Cincinnati for her helpful assistance. Paul Cocca of the USEPA has also assisted with helpful discussion about the use of BASINS.
References Abbott, M. B., Bathurst, J. C., Cunge, J. A., O'Connell, P. E. and Rasmussen, J. (1986). An introduction to the European Hydrological SystemÐSysteme Hydrologique Europeen, SHE, 1. History and philosophy of a physically-based, distributed modeling system. Journal of Hydrology 87, 45±59. Arnold, J. G., Williams, J. R., Nicks, A. D. and Sammons, N. B. (1990). SWRRB, A Basin Scale Simulation Model for Soil and Water Resources Management. College Station: Texas A & M University Press. Arnold, J. G., Williams, J. R., Srinivasan, R., King, K. W. and Griggs, R. H. (1994). SWATÐSoil and Water Assessment ToolÐUser Manual. Agricultural Research Service, Grassland, Soil and Water Research Lab, US Department of Agriculture. Beasley, D. B., Huggins, L. F. and Monke, E. J. (1980). ANSWERS: A model for watershed planning. Transactions of the American Society of Agricultural Engineers 23(4), 938±944. Bhaduri, B., Grove, M., Lowry, C. and Harbor, J. (1997). Assessing long-term hydrologic effects of land use change. Journal of the American Water Works Association 89(11), 94±106. Bhaduri, B., Minner, M., Tatalovich, S. and Harbor, J. (2001). Long-term hydrologic impact of urbanization: A tale of two models. Journal of Water Resources Planning and Management 127(1), 13±19. Bicknell, B. R., Imhoff, J. C., Kittle, J. L. Jr, Donigian, A. S. Jr, Johanson, R. C. (1996). Hydrological Simulation ProgramÐFortran User's Manual
S. T. Y. Tong and W. Chen
for Release 11. Athens, GA: Of®ce of Research and Development, U.S. Environmental Protection Agency. Bolstad, P. V. and Swank, W. T. (1997). Cumulative impacts of land use on water quality in a southern Appalachian watershed. Journal of the American Water Resources Association 33(3), 519±534. Bouraoui, F., Vachaud, G. and Chen. T. (1998). Prediction of the effect of climatic changes and land use management on water resources. Physics and Chemistry of the Earth 23(4), 379±384. Burnash, R. J., Ferral, R. L. and McQuire, R. A. (1973). A Generalized Stream Flow Simulation System in Conceptual Modeling for Digital Computers. Sacramento, CA: US National Weather Services. Changnon, S. A. and Demissie, M. (1996). Detection of changes in stream¯ow and ¯oods resulting from climate ¯uctuations and land use drainage changes. Climatic Change 32, 411±421. Debrewer, L. M., Rowe, G. L., Reutter, D. C., Moore, R. C., Hambrook, J. A. and Baker, N. T. (2000). Environmental Setting and Effects on Water Quality in the Great and Little Miami River Basins, Ohio and Indiana. National Water-Quality Assessment Program. Water-Resources Investigations Report 99-4201. Donigian, A. S. Jr, Imhoff, J. C., Bicknell, B. R. and Kittle, J. L. Jr (1984). Application Guide for the Hydrological Simulation ProgramÐFORTRAN. Athens, GA: Environmental Research Laboratory, US Environmental Protection Agency. Donigian, A. S. Jr. and Huber, W. C. (1991). Modeling of Nonpoint Source Water Quality in Urban and Nonurban Areas. Athens, GA: Environmental Research Laboratory, US Environmental Protection Agency. Ferrier, R. C., Whitenhead, P. G., Sefton, C., Edwards, A. C. and Pugh, K. (1995). Modelling impacts of land use change and climate change on nitratenitrogen in the River Don, North East Scotland. Water Research 29(8), 950±1956. Fisher, D. S., Steiner, J. L., Wilkinson, S. R. (2000). The relationship of land use practices to surface water quality in the Upper Oconee Watershed of Georgia. Forest Ecology and Management 128, 39±48. Gburek, W. J. and Folmar, G. J. (1999). Flow and chemical contributions to stream¯ow in an upland watershed: a base¯ow survey. Journal of Hydrology 217, 1±18. Hanratty, M. P. and Stefan, H. G. (1998). Simulating climate change effects in a Minnesota agricultural watershed. Journal of Environmental Quality 27, 1524±1532. Henderson-Sellers, A. (1994). Land-use change and climate. Land Degradation Rehabilitation 5, 107±126. Hulme, M., Hossell, J. E. and Parry, M. L. (1993). Future climate change and land use in the United Kingdom. Geographical Journal 159(2), 131±147. Johanson, R. C., Imhoff, J. C., Kittle, J. L. Jr and Donigian, A. S. (1984). Hydrological Simulation ProgramÐFORTRAN (HSPF): Users Manual for Release 80. Athens, GA: Environmental Research Laboratory, US Environmental Protection Agency. Knisel, W. G. (1980). CREAMS. A Field Scale Model for Chemicals, Runoff and Erosion from Agricultural Management Systems. US Department of Agricultural Conservation Research Report, 26. Krysanova, V. and Luik, H. (1989). Simulation Modeling of a System Watershed-river-sea Bay. Tallinn, Valgus.
Krysanova, V., M uller-Wohlfeil, D.-T. and Becker, A. (1998). Development and test of a spatially distributed hydrological/water quality model for mesoscale watersheds. Ecological Modelling 106, 261±289. LeBlanc, R. T., Brown, R. D. and FitzGibbon, J. E. (1997). Modeling the effects of land use change on the water temperature in unregulated urban streams. Journal of Environmental Management 49, 445±469. Lenat, D. R. and Crawford, J. K. (1994). Effects of land use on water quality and aquatic biota of three North Carolina Piedmont streams. Hydrobiologia 294(3), 185±200. Leonard, R. A., Knisel, W. G. and Still, D. A. (1987). GLEAMS: Groundwater loading effects on agricultural management systems. Transactions of the American Society of Agricultural Engineers 30(5), 1403±1428. Lerch, N. K., Hale, W. F. and Milliron, E. L. (1975). Soil Survey of Clermont County, Ohio. Department of Agriculture. Liu, A. J., Tong, S. T. Y. and Goodrich, J. A. (2000). Land use as a mitigation strategy for the water-quality impacts of global warming: A scenario analysis on two watersheds in the Ohio River Basin. Environmental Engineering and Policy 2, 65±76. Mander, U., Kull, A., Tamm, V., Kuusemets, V. and Karjus, R. (1998). Impact of climatic ¯uctuations and land use change on runoff and nutrient losses in rural landscape. Landscape and Urban Planning 41, 229±238. Mattikalli, N. M. and Richards, K. S. (1996). Estimation of surface water quality changes in response to land use change: application of the export coef®cient model using remote sensing and geographical information system. Journal of Environmental Management 48, 263±282. Meissner, R., Seeger, J., Rupp, H. and Balla, H. (1999). Assessing the impacts of agricultural land use changes on water quality. Water Science Technology 40(2), 1±10. Nikolaidis, N. P., Heng, H., Semagin, R. and Clausen, J. C. (1998). Non-linear response of a mixed land use watershed to nitrogen loading. Agriculture, Ecosystem and Environment 67, 251±265. Ohio Environmental Protection Agency. (1996). Ohio Water Resource Inventory, Executive Summary: Summary, Conclusions, and Recommendations. Columbus, OH: Division of Surface Water and Monitoring Assessment Section. Ohio Department of Natural Resources. (1964). Water Inventory of the Little Miami and Mill Creek Basins and Adjacent Ohio River Tributaries. Ohio Water Plan Inventory, Report 18. Rai, S. E. and Sharma, E. (1998). Comparative assessment of runoff characteristics under different land use patterns within a Himalayan watershed. Hydrological Processes 12, 2235±2248. Singh, V. P. (1995). Environmental Hydrology. Borton, MA: Kluwer Academic. Syvitski, J. P., Morehead, M. D. and Nicholson, M. (1988). HYDROTREND: A climate-driven hydrologictransport model for predicting discharge and sediment load to lakes or oceans. Computers and Geosciences 24(1), 1±110. Tong, S. T. Y. (1990). The hydrologic effects of urban land use: A case study of the Little Miami River Basin. Landscape & Urban Planning 19, 99±105.
Land use and surface water quality Tsihrintzis, V. A. and Hamid, R. (1998). Runoff quality prediction from small urban catchments using SWMM. Hydrological Processes 12, 311±329. US Environmental Protection Agency. (1984). Users Manual for Hydrological Simulation ProgramÐ FORTRAN (HSPF). Athens, GA: Environmental Research Laboratory, EPA-600/3-84-066. US Environmental Protection Agency. (1998). Better Assessment Science Integrating Point and Nonpoint Sources. Of®ce of Water. EPA-823-B-98-006. US Environmental Protection Agency. (2000). BASINS Technical Note 6: Estimating Hydrology and
Hydraulic Parameters for HSPF. Of®ce of Water. EPA-823-R00-012. Wu, R. S. and Haith, D. A. (1993). Land use, climate, and water supply. Journal of Water Resources Planning Management 119(6), 685±704. Young, R. A., Onstad, C. A., Bosch, D. D. and Anderson, W. P. (1989). AGNPS: A non-point source pollution model for evaluating agricultural watersheds. Journal of Soil and Water Conservation 44(2), 168±173.