First Draft

Ph. D Research Proposal

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now


Israel Olaoye?
Coastal wetlands and estuaries are transition zones for the transport of sediments from source to depositional areas. They function as the storage for the pollutants, nutrients and sediments nutrients being washed from the watershed (Biggs and Howell, 1984). The Old Woman Creek (OWC) estuary is a natural coastal wetland in the shore of Lake Erie which supplies drinking water to coastal communities in the Northeastern Ohio (Ortiz 2013). The OWC performs many environmental functions which include flood control, water quality regulation and natural habitat for aquatic and wild life (Herdendorf et al., 2006). OWC estuarine extends to the south of Lake Erie for about 1.3 mi and has an area of about 150 acres (Fig 1&2). Its greatest width is 0.21 mi and average depth is about 1.5 ft. Land use changes in the watershed such as agriculture and industrialization often increase erosion and sedimentation rate in the wetland thereby decreasing its effectiveness in capturing and storing contaminants (Evans and Seamon, 1997). Over the years, there has been a rising sea level and increased sedimentation rate into OWC estuary and yet it is not filled up, this has prompted more research in OWC (Reeder and Eisner, 1994 and Krieger, 2002). This study will therefore focus on how changes in hydrology and hydrodynamics will influence the sediment transport and retention in the estuary (wetland) portion of Old woman creek.
The general rise in population and the changes in socio economic factors has necessitated changes in land use/land cover of various regions. The study of changes in land use/land cover of an area is important in understanding past events as well as predicting future conditions (Turnel et al., 2001). Land cover includes all the physical material on the earth’s surface , examples are water, asphalt, bareground and trees while Land use describes the various way land is used by human beings, examples are agricultural and urban land use (Fisher, Comber, ; Wadsworth, 2005). Land use/land cover modify the environment in ways which affects humans’ living conditions. Some of the results of land use land cover changes include, loss of natural environments and biodiversity, and increased in the concentration of greenhouse gases (Turnel, 2001)
Ecologists are particularly concerned about the effect of land use change on ecosystems and biodiversity (Turnel, 2001). Changes in land use patterns in a catchment can have great implications on availability of water supply and water quality. For instance, urbanization is associated with increased surface runoff, reduction in groundwater recharge and increased concentration of loads. The effect of land use/land cover changes observed in a watershed assist in land use planning and water management (Amna B. et. al., 2015)
Remote Sensing is a powerful tool for deriving accurate and timely information on the spatial distribution of land use/land (Carlson T.N., Azofeifa S.G.A., 1999). Landsat images have contributed a great deal to the classification of different landscape components (Ozesmi and Bauer, 2002). Landsat images can be used to detect changes in features of interest on digital images acquired in different years. Many techniques to interpret satellite imageries have been developed using post classification comparison, conventional image differentiation, image ratios and image regression (Lu D et al., 2005). Post classification comparison involves grouping the land use/land cover of an area into major classes using remote sensing or GIS tool and comparing with older classes to detect changes. Some studies have found post classification comparison to be the most accurate procedure (Mas J. F 1999 and Yuan F et. al., 2005). In this study, post classification comparison will be carried out on Landsat 5 and 8 images of Old Woman Creek to understand the land use/land cover changes from 1993 to 20011.
Principal Component Analysis (PCA) is “the general name for a technique which uses sophisticated underlying mathematical principles to transforms a number of possibly correlated variables into a smaller number of variables called principal components”(Richardson, 2009). PCA is employed to reduce data sets dimensionality and it uses a method of vector space transform. The variables of a data set can be simplified to few variables (principal components) by mathematical projection which would allow patterns, trends and outliers in the data to be noted easily, thereby enhancing reasonable interpretation (Richardson, 2009).
Bryant and Yarnold (1995, p. 132) define Rotation as “a procedure in which the eigenvectors (factors) are rotated in an attempt to achieve simple structure.” Varimax Rotation is a reversal of coordinates method of analysis that optimizes the variance sum of the loadings in principal component analysis (PCA) and producing either large or near zero coefficients(“Tanagra – Data Mining and Data Science Tutorials: VARIMAX rotation in Principal Component Analysis,” n.d.). In PCA, the direction in which variation takes place is the orthogonal eigenvectors while the magnitude of such variability is the eigenvalues. The eigenvector with maximum eigenvalue is the principal components. Kaiser 1958 developed Varimax rotation which enhances data variability by ensuring variance optimization on the all the principal components while keeping constant their orthogonal direction. The relationship between the VPCA components and the data points is computed from the input data and VPCA results and is termed component scores. The VPCA components of a reflectance dataset is identified by comparing the spectral signature of the components to some known or standard spectra (Ortiz, 2013). In this work, VPCA will be carried out on Landsat 5 and 8 imageries of the OWC while isolating the land pixel to determine the constituent materials in the estuarine water
Hydrology, the study of the occurrences, properties, circulation and distribution and relationship of water to living things is very important as it helps to understand various changes in the hydrologic cycle brought about by changes in land use and urbanization (Ray 1975). A hydrological model is a representation of a real world hydrological system used for predicting the system performance and understanding the processes involved. The best model is simple, requires least input parameters and gives close results to reality while predicting a catchment. The basic input for all models are rainfall data and drainage area, other data are vegetation cover, land use/land cover, topography soil moisture, soil properties and aquifer properties (Sharma et al., 2008). Soil and Water Assessment Tool (SWAT) is a physically based watershed scale complex model designed to simulate the effect of land management on water flow and sediment circulation and agricultural yield with chemicals in any basin either gaged or ungaged (Documentation, 2009 and Devi. Ganasri, &Dwarakish, 2015). SWAT model will be used in this work to simulate water discharge, sediment flow (erosion) and nutrient flow at OWC.
“Turbidity is the measurement of scattered light that results from the interaction between a beam of light and particulate material in a liquid sample. It is an expression of the optical properties of a sample that causes these light rays to be scattered and absorbed rather than transmitted in straight lines through the sample”. (Sadar, 2009). Particles dissolving in water affecting the color of water and thus, the turbidity measurement. (U.S. Geological Survey TWRI Book 9 Version 2.1 9/2005). Turbidity measurement is qualitative but its comparison with known standards makes it quantitative. The fundamental standard for turbidity measurement is farmazin which is used for the calibration of many turbidimeters (Sadar, 2010). Turbidity data can be used for the following; real- time water monitoring to understand watershed conditions, monitoring water quality for aquatic life, maintaining the quality of public drinking water, finding out the impact of land use and various human activities on water (Gray and Glysson 2003) and this is the reason for turbidity measurement in this study. In short, this work involves the use of land use/land cover classification method of remote sensing, VPCA statistical method, turbidity measurement and hydrological modeling to understand the effect of climate change and human activities on OWC, quantify the nutrient, sediment and water transport, relate the changes in turbidity to the changes in land use and simulate the discharge and the effect of flash flood on the OWC.

Figure 1. Location map of Old Woman Creek estuary and watershed (Herdendorf et al., 2006)

Fig 2. View of Old Woman Creek (GoogleEarthPro)

Justification for the study (references listed)
The research and monitoring program at OWC is established to improve our knowledge of the Great Lakes estuarine systems and understand the governing conditions of the OWC estuary. The essence of this program is to provide information to the OWC governing bodies, resource agencies and private sectors interested in coastal management and development. This will assist them in making good decisions regarding land use, waterway management, and development. The long-term objective of the research program is to develop a broad knowledge about the fresh water estuarine system from the results of all the projects on OWC (Herdendorf et. al. 2006).
Sediment transport and turbidity studies have been modeled in the watershed portion of Old Woman Creek by Wilson et al at Case Western University, but not the estuary and, as such, less information is available about sediment transport in the estuary and the associated wetlands of OWC. Thus, in addition to studying the effect of land use/land cover on the water of the OWC and studying the water constituents, this work attempts to examine the hydrologic and hydrodynamic response to changes in storms and water levels within the estuary. This will be achieved by developing a hydrological model for the estuary and simulating storm events of different magnitude. Water input and hydrological fluxes will be studied to understand, predict, and manage the water resources of OWC. In addition to the above, this work will develop a satellite remote sensing monitoring program for Old Woman Creek through the analysis of satellite imageries.
Site Location and Character:
Old woman Creek National Estuarine Research Reserve is located on Latitude 41? 23’N and Longitude 82? 33’W which falls on the eastern part of Huron city and southern part of Lake Erie. Major land use in OWC watershed is Agriculture which is responsible for the pollutants in OWC estuary. The tidal range in Lake Erie and OWC is about 4cm and the average water salinity is about 1ppt. There is daily, seasonal and annual variability in water level in the creek and the estuary and this is caused by the storm runoff into the OWC, the changing water level in Lake Erie, seiches on the lake and the opening and closing of the mouth throughout the year. (“Old Woman Creek,” n.d.)
Barrier Beach Status and Water Exchange
There is surface water exchange between Lake Erie and the estuary because of the opening and closing of the mouth. When the mouth is open, wind drives water to flow into the estuary and out of the estuary repeatedly and this affects the water quality in the estuary, thus the turbidity data at the OL and WM stations are influenced by the opening and closing of the mouth. Storm event or precipitation can open the barrier beach rapidly while closing of the beach occurs gradually with decrease in water level (

Objectives of the study
The main objective of this research is to develop a climate change monitoring scheme for the OWC estuary. This will involve using remote sensing techniques and GIS to evaluate the impact of physical changes in the vicinity of OWC on water and the structure of the estuary. I will determine the principal components of the material in the estuary water through analysis of satellite imageries and account for their source. I will evaluate the impact of land use/land cover changes on the OWC and the quality of the estuarine water. I will establish possible relationship between land use, water quality, erosion, sedimentation and nutrients accumulation. Lastly, this study will assist in predicting the influence of changes in hydrology (due to different precipitation patterns) on OWC using a hydrological model.

The above objectives will be achieved through the following tasks
Model land use land cover change of the OWC from 1983 to 2011 using Landsat 5 imageries through supervised and unsupervised classification done on TerrSet Interface
Carry out VPCA on seven Landsat 8 images using IDL (8.6.1)- ENVI(5.4.1) interface to determine the principal components material in the estuary water.
Interpret the 2002- 2017 turbidity data acquired using Sondes at 4 gage stations in OWC to determine the quality of the estuary water and relate turbidity to land use/ land cover
Develop a watershed scale hydrological model for the OWC using SWAT model and simulate discharges, nutrients, sediment and flood events of different magnitude for previous and future years
Relate turbidity to VPCA results.
4.0 Data Sources and software (references listed)
Land use/ Land cover analysis: Data includes Landsat -5TM and Landsat 8 imageries available free of charge at earthexplorer ( The TM satellite visits a location on the earth’s surface every 16 days around noon (Eastern Time) and collects reflectance data in Electromagnetic (EM) spectrum using the visible, near infrared and mid infrared regions only. (Table 1.0). The resolution of the data is 30 m. I registered on the web page and downloaded data using Path19 Row 31. Software packages include TerrSet Geospatial Monitoring and modeling system for remote sensing, available in Room 403 laboratory. Data was imported into Terrset, zoomed and clipped to focus on OWC and its environs for analysis
VPCA: Landsat 5 and Landsat 8 imageries downloaded from earthexplorer will be processed using ENVI 5.4.1-IDL 8.6.1 interface and StatFi on Microsoft excel
3.Turbidity Data: Time series turbidity data taken every 15 minutes at 4 gaging stations in OWC is available. Three of the logging stations are within the estuary: first at Darrow Road (DR), second is near State Soute 6 (WM), third is upstream side of WM station around the lower estuary (OL). The forth is station (BR) is situated at the upstream side above the estuary. The OL station helps to understand the influx of lake water into OWC during seiche events. (Fig 3). Discharge data is available from 2007 through 2017. Turbidity data is available from 2002 to 2017
Hydrological Model: Data include; Light detection and ranging (LiDAR) data for OWC watershed area is obtained from Ohio Geographically referenced information program (OGRIP) as the bare earth digital elevation model (DEM). Land cover data for OWC is obtained from National Land Cover Database (“MRLC NLCD 2011,” n.d.). Soil data is downloaded from STATSGO (“Description of STATSGO2 Database | NRCS Soils,” n.d.). Digital soil map of the world was downloaded from FAO food and Agriculture website. Climate and Stream discharge data were downloaded from USGS national water information system. I am using SWAT model which runs on ArcGIS interface called ArcSWAT. I downloaded it from (“ArcSWAT | Soil and Water Assessment Tool,” n.d.) and run it on ArcGIS 10.4.1.

Fig. 3. Gage stations and available data

Table 1.0: Landsat 5 TM Band Definitions (Obenschain et al., 1997)
Band Type Wavelength (µm) Resolution (m)
1 Blue-Green 0.45-0.515 30
2 Green 0.52-0.605 30
3 Red 0.63-0.69 30
4 NearIR 0.77-0.90 30
5 SWIR 1.55-1.75 30
6 SWIR 10.4-12.5 30
7 SWIR 2.09-2.35 30

Remote Sensing
Quickbird and Landsat TM
Quickbird acquires very high-resolution multispectral images in the visible, near infrared and panchromatic bands which are capable of giving detailed information on the water quality of OWC. The spectral resolution of the bands is 450-520nm for blue, 520- 600 nm for green, 630-690 nm for red, 760-900 nm for NIR and 450-900 nm for panchromatic(“QuickBird | Satellites | Geoimage,” n.d.). QB images for OWC will be acquired at two different dates in the summer (June and October 2011) for detailed water quality analysis and the results will be compared with the Thematic Mapper images taken over the same time span. The data will be adjusted based on the spectral resolution of QB (2.6 m). This will ensure compatibility as well as comparisons with future data and increase the signal to noise ratio without interfering with the current resolution of QB data. Water quality data will be available from the high resolution multispectral QB images and the variation of water quality with time within the Old Woman Creek estuary from June to October 2011 will then be established. I will then apply the spectral decomposition method that gives results across different pixel sizes (Ortiz et al., 2013 and Lekki et a1., 2017)
Landsat 5 TM images acquired within similar time will be processed using Terrset (“TerrSet Geospatial Monitoring and Modeling Software,” n.d.). Atmospheric correction and radiometric calibration will be carried out on the images to remove atmospheric effect and scattering. The wavelength will be used to calculate the derivative reflectance (Demetriades-Shah et al., 1990).
Green Red Derivative = (Red- Green)/ (Red mid-Green mid).
NDVI= (NIR-R)/(NIR+R) and NDWI= (NIR-B)/(NIR+B).

Land Cover and Land use type in the vicinity of Old Woman Creek.
Supervised and Unsupervised Classification using TerraSet Software
1983 and 2011 Landsat-5 TM data, Path 19 Row 31 imagery of the Old Woman Creek was imported into TerrSet software for analysis. The data was downloaded from USGS earthexplorer extracted into a project folder and imported into TerraSet software. In TerraSet, the images were zoomed in to focus on the Old Woman Creek estuary and its environs. True and False color composite images were created to make visual identification of the features in the image. Images were digitized and classified into various land use and land cover types using the tools provided in the TerrasSet software and about 28 different land use/land cover types were first identified. By visual comparison of the image with the google earth satellite imagery of OWC, we established that the major land use/land cover in OWC are water, built environment, wetland and vegetation on land
Supervised Classification
Images were digitized in TerrSet and training areas were selected to help the algorithm define the basis for classification. Training areas are regions which serves as examples for each class. Spectra Signature was assigned on the training areas, this helps the software identify the class which bears the greatest resemblance to the training areas. Maximum likelihood classification was run on the images. This algorithm assigns pixels to class based on probability density function. Supervised classification gives better results but require field information for processing

Unsupervised classification
Usually used as a preliminary classification method to study the relationship between different land covers and spectral signatures (German 2014). This is performed by running the Cluster classifier tool from Idris Image processing via Hard classifier. This algorithm uses statistical method to group signature into spectral classes.
Land Cover Modeler: Land cover modeler was run on both the 1988 and 2011 images in order to quantify the changes in land use within the vicinity of OWC over time.
The results of supervised and unsupervised classification were compared to establish the land use classes. The results of land cover modeler will both 1988 and 2011 were compared to detect changes in land use/ land cover in OWC overtime. Bar charts and graphs were derived from the land cover modeler to depict these changes quantitatively.
Normalize Difference Water Index (NDWI)
The NDWI of blue and middle infrared (MIR) bands (NDWI1;5) was used to identify the open-water regions. One normalized difference remote sensing index was investigated using the visible and infrared Landsat-5 TM bands. NDWI identifies water bearing pixels based on the ratio of the difference between the visible and the infrared band over the sum of the same with the aim of identifying open water

The following formula was used to calculate the NDWI:

NDWI= Rband1- Rband5
Rband1 + Rband5

R is the percent reflectance of the band denoted as the subscript. Blue band is 1. (0.45 -0.52 ?m) and the middle infrared is 5 (1.55-1.75 ?m). The index is a ratio transform which contrasts the energy reflectance from open-water with that of other land covers such as urban. The index value ranges from -1 to +1.
Results of Remote sensing analysis
Unsupervised Classification
The results of unsupervised classification are shown in figures 4;4. Both 1983 and 2011 results show three major classes of land cover within the vicinity of Old Woman Creek which are vegetation, water, and road and buildings. In 1983 image, the estuary has a larger size with wider wetland and greater water influence. There is more vegetation in the watershed and less or no vegetation within the wetland (estuary) and the entire wetland is filled with water. There are also some pounds of water scattered within the watershed which also contribute to the volume of water in the estuary. Development was low as seen in the buildings and roads. In 2011 however, the size and geometry of the watershed has reduced considerably and much of the estuary water has been lost. Much of the wetland area is now dominated by vegetation as opposed to the sparse vegetation in 1983. Also, building and roads have occupied most of the previously vegetated areas of the watershed.
Supervised Classification
The results of supervised classification are shown in figures 6;7. Four major classes of land cover were identified which include water, vegetation, land with sparse vegetation and Built environment. In the 1983 image, the geometry of the estuary (wetland) was well defined and the star highland was located within the estuary. Much of the watershed is dominated by vegetation and dry land. Vegetation is scattered all round the vicinity of the Old woman creek. In 2011 however, the geometry as well as water volume in the estuary has reduced drastically due to sediment influx and much of the wetland is dominated by vegetation. Vegetation has also grown in the previous dry land around the wetland. Buildings and roads have occupied more part of the watershed.
Land Cover Modeler: Land cover modeler was run on the supervised classification to obtain various percentage changes is land cover. It was estimated that the percentage increase in buildings and roads from 1983 to 2011 is 56.4%. The percentage increase in vegetation from 1983 to 2011 is 32. 4%. Figure 8 is the map showing the gain and loss in buildings and roads in Old woman creek area from 1983 to 2011 with more gain than loss. Figure 9 shows the gain and loss in vegetation while figure 10 shows considerable loss in water within the period in question. Figure 11 shows all changes in the old woman creek area within the period in question. These results corroborate the results obtained from the NDWI map for the years1983 and 2011 (Figures 12;13).
The results presented above shows that there has been a significant change in the size, water volume and geometry of the Old woman creek estuary overtime. Land use type in the in the vicinity of the Old Woman Creek estuary has also changed overtime and this has affected the estuary. A vast area of the watershed is now used for agricultural purposes. The use of fertilizer for agricultural purposes has brought about increase in the influx of nutrients into the estuary and this has led to increased vegetation growth within the watershed as shown by the images presented. The growth of vegetation in the estuary is accompanied by the growth of algae which normally blooms on the surface.
Moreover, drastic reduction in the size and geometry of the old woman creek estuary as at 2011 compared to 1983 could be attributed to the effect of climate change. Increase in surface temperature brings about longer increase in evapotranspiration could contribute to loss of water in the estuary. Additionally, influx of sediment into the estuary can reduce the size and depth of the estuary thereby pushing water from the estuary into Lake Erie. The result of this findings corroborates the findings of Wijekoon, 2007.

Figure 4: Unsupervised classification (1983) Fig 5: Unsupervised Classification (2011)

Fig. 6: Supervised classification (1983)

Figure 10. NDWI (1983) difference map

Figure 11. NDWI (2011) difference map


Principal Components Analysis (PCA)
Data Preparation and processing
Landsat imageries downloaded from USGS earthexplorer are uploaded into ENVI/IDL software to prepare the data for PCA decomposition. ENVI works similar to Terrset and provides graphics and editing functions for manipulating imageries to the desired sizes and colour. IDL is a coding platform which allows different codes to be written and stored for various analysis on the imageries. Imageries were first treated on the ENVI interface. The imageries were zoomed and clipped to the region of interest (OWC) in a rectangular shape. Land pixels were masked and then removed from the imageries to eliminate the effect of land pixels thereby increasing the signal to noise ratio. The “noland” imageries are exported into the IDL interface where the derivative imageries were created, and all imageries stacked for PCA. PCA is done in ENVI interface and the results exported into IDL for VPCA
VPCA Decomposition of Derivative Spectra
VPCA decomposition can be done on both the field data and satellite imageries. Remote sensing imageries are 3-dimensional in nature with x- dimension being the longitude, the y-dimension being the latitude and z-dimension being the wavelength of the reflectance data. PCA results were exported to IDL where VPCA decomposition was carried out using algorithm developed by Dulci Avouris. VPCA scores, VPCA flips and the variance were extracted and were further processed using StatsFi in Microsoft excel to identify the VPCA component. The VPCA components are plotted and compared with the spectral library to find out the water components ( Figures 14, 15;16).

Fig 14: Determination of Chlorophyta component

Fig 15: Determination of Cyanobacteria component

Fig 16: Determination of Haematite component

Turbidity Measurements and Interpretation
Turbidity is mostly measured in Nephelometric Turbidity Unit (NTU), or the equivalent Formazin Nephelometric Unit (FNU). Nephelometry involves directing a stream of light to a liquid sample and measuring the intensity of light scattered at 900 to the light stream. The NTU/FNU scale was formed with reference to the suspension of formazin particles. (Kelly et al., 2014). Turbidity devices developed recently use photosensitive cells to find the amount of transmitted and scattered light. The laboratory equipment has five main parts: a light source (b) Lens (c) Photosensors (d) sample cell and (e) logger Fig 17 (Lawler, 2005). The 2016 monitoring program collected data at the four stations in OWC from March to December 2016 using the YSI EXO2 sonde after which all sondes were replaced by a similar sonde, YSI 6600V2 datalogger for data collection till date. Data available are temperature, conductivity, salinity, turbidy and pH at a water level or depth. The EXO2 measures turbidity in formazin nephelometric units (FNU) while 6600V2 datalogger measures turbidity in nephelometric turbidity units (NTU) but the two units are essentially the same. These loggers simultaneous measure depth or water level. Data analysis and graphing will be done in Microsoft excel. I will make time series trend analysis for high turbidity days (number of days with turbidity ? 5NTU) to find out how high turbidity days varies with time. I will make seasonally-adjusted trend analysis of turbidity data to study the long-term change in turbidity. I will determine the low turbidity status, high turbidity status and occasional spikes and will discuss on the turbidity range of OWC overtime. “Low Turbidity Status is the turbidity baseline lower than 5 NTU greater than or equal to 50% of the time. High Turbidity Status is the turbidity baseline lower than 5 NTU less than 50% of the time. Occasional Spikes is the sudden increases in turbidity that exceed 5 NTU occurring less than 30 times/year. Frequent Spikes is the sudden increases in turbidity that exceed 5 NTU occurring more than 30 times/year ” (Quality, Range, Water, ; Areas, 2010) (Figures 18;19)


Fig18. Turbidity in Old Woman Creek

Fig 19. Highest raw turbidity recorded at Old Woman Creek (2002-2017)

SWAT Model
In SWAT model, the watershed may be divided into sub basins when the land use and soil types are highly variable and can affect the hydrology, this will allow for spatial references of different areas within the watershed. Input data for each sub basin is assembled as climates, hydrological response units (HRUs), ponds/wetlands and groundwater. Different HRU is a combination of a unique soil type, landcover type and management practice. (Documentation, 2009). SWAT model can simulate water discharge and sediment circulation, nutrient and pesticide circulation from input data which are daily rainfall data, maximum and minimum air temperature, wind speed, solar radiation and relative humidity. Evapotranspiration is estimated using Penman Monteith, Priestly- Taylor and Hargreaves methods (Devi, Ganasri, ; Dwarakish, 2015).
Since water balance drives all hydrological processes within the watershed, all accurate simulations must use all water circulation and must bear close resemblance to the natural system in a watershed(Devi et al., 2015) and (Documentation, 2009). The hydrology of a watershed is divided into the land phase and water phase. The land phase regulates water inflow, sediment, pesticides and nutrient transport into the different sub watersheds while the water phase moves sediment, nutrients and water from the inlet to the outlet of the watershed (fig 20;21). (Documentation, 2009)

Fig 20. Schematic representation of the hydrological cycle (SWAT Documentation, 2009)

The water balance equation used in SWAT simulation is shown below
?SW?_t=SW_0+ ?_(t=1)^t??( P- Q_surf- E_a ?-w_seep-Q_gw)

Where SWt is the soil humidity, SW0 is the initial humidity, P is precipitation in mm, Q is the runoff on the surface. Ea is the evapotranspiration, wseep is the water seepage from the subsurface soil and Qgw is the base flow from groundwater and t is the time in days(Documentation, 2009)

Fig 21. HRU in a command loop (SWAT Documentation 2009)

Hydrological Modeling of Old Woman Creek
ArcSWAT is a GIS interface on which SWAT model is run. It was downloaded from SWAT website and installed on ArcGIS 10.4. The date and time setting of the computer was adjusted to US only in the format MM/DD/YYYY to enable SWAT run. The stages involved in the running of SWAT model include the following
Data preparation: This is the most cumbersome stage in SWAT modeling, it requires downloading of data, preparation of maps and tables. The data to be prepared for OWC are
Digital Elevation Model: To fully estimate the extent of old woman creek watershed, four 30 m DEMs covering Erie, Lorain, Huron, Ashland, Crawford, Richford and Seneca counties were downloaded from the USGS national map site, extracted and imported into ArcGIS, and Musaic to form a single composite DEM. I identified the UTM zone for OWC to be 17N by applying the UTM shape file from the reference system to the DEM and re-projected the DEM to WGS 1984 UTM 17 N. This makes the DEM ready to be used in SWAT model Figurs 22;23

Land Use/land Cover map: The boundary shape file of the 7 counties listed above were extracted from US counties shape file. These were jointed into a single composite shape file by the dissolve tool in ArcGIS (Fig 24). US Land cover map was downloaded from National Land Cover Database 2011 (NLCD 2011). It was classified into four major land use/land cover which are Urban, Water, Agriculture/vegetation and wetland and clipped to composite shapefile of the study area earlier created. The land use/ land cover map was placed on the DEM and the DEM cut into the shape of the study area (fig 25;26). Textfile was created for the land use/land cover map (Table 2) to enable SWAT to read the various land use/land cover in codes and the map was ready for SWAT model.
Soil Data: The shapefile of the world soil data was downloaded from Fao digital map and was imported into ArcGIS. This map was converted to raster and clipped to the shapefile of the study area (Fig 27).
Weather Data: Weather data for the study area was downloaded from Global Weather data for SWAT (“Global Weather Data for SWAT,” n.d.), The data is from 1979-2014. It was edited and prepared for SWAT using notepad++
Setting up SWAT Project Folder: SWAT project folder is set up in the main drive of the computer or as a fist level folder for SWAT to run. SWAT will not run if the project folder is not a first level folder. In this work, SWAT project folder was set up in this format: “D: SWATProject”.
Project Setting- watershed and sub-watershed delineation: ArcSWAT was opened and the SWAT tools activated. A new project was set up for SWAT and was named. Prepared DEM and land cover data were added to this project and watershed was delineated using automatic watershed delineator through a series of tasks contained in the watershed delineator dialogue box. These tasks include; DEM set up, Stream definition, flow direction and accumulation, Minimum working area (a minimum area of 500 Ha is required for small watershed, Stream network (“create stream and outlets”). Then the stream network with various outlets and inlets was created (Fig 28). The many outlets generated were removed and a major outlet for the watershed was fixed manually and watershed delineated (Fig. 29, 30 ; 31). Then sub-basin parameters were calculated, and watershed report generated.
Hydrologic Response Units (HRU) Analysis: to define my HRUs, I will assign the percentages to land use, soil class and slope class over the sub-basin area and create multiple HRUs. SWAT model automatically generates several HRUs based on the input parameters.
Weather data definition: I will load the corrected Global weather data and specify “daily” timestep for precipitation. I will load the index files for both precipitation and temperature and write input tables for all the hydrological components and update the database.
SWAT Run: I will run SWAT model, specify warm up years to be 3, simulate discharge for a period of 10 yeas (2004-2014) and generate output daily. I will include soil nutrient, water quality and pesticides in the simulation and generate diagrams for the results in the simulation. I will read all output data from the database and plot the (flow-output) discharge data as hydrographs. I will extract information about sediment input and sediment output
Manual Calibration: I will carry out manual calibration with a manual calibration tool provided in SWAT, run a SWATcheck to see the effect of the calibration and run SWAT again to generate a new set of output. I will plot the hydrograph of the discharge data and compare the discharge from calibrated to the discharge from uncalibrated output by calculating the correlation coefficient r2
Simulation into the future: SWAT is able to simulate into the future using the average climate data for the entire climate years. I will define in SWAT all the weather information including rainfall, temperature, relative humidity, solar radiation, winspeed. I will rewrite SWAT tables and update the database and then run SWAT simulation.
In all, sections a-f have been done and the results are presented as figures during discussion

Future Work
There are a lot of work to be done in each method as discussed here.
– QB method of classification will be done and comparison will be made with remote sensing method.
– Turbidity data will be processed as discussed to understand the variation in turbidity with time, account for such variation and relate turbidity to with land use.
– Complete work on VPCA on Landsat 5;8 images
– Hydrological model of the Old Woman Creek will be completed and discharges, nutrients, sediment will be simulated. Flood will be also be simulated in the watershed.
– Relate turbidity algorithm to SWMP sensors
– Relate turbidity to VPCA


I'm James!

Would you like to get a custom essay? How about receiving a customized one?

Check it out