Landscape Modeling

Principal Investigator
Brian Lee, Department of Landscape Architecture

Project Summary
A land use modeling study in the Lexington-Fayette Area of Central Kentucky

Project Description
The purpose of this project is to implement the SLEUTH cellular automaton urban simulation model in central Kentucky. SLEUTH is an acronym for Slope, Land cover, Exclusion, Urbanization, Transportation, and Hillshade. SLEUTH was initially developed by Dr. Keith Clarke of the University of California - Santa Barbara. A cellular automaton is a collection of cells on a grid that changes states (developed or not developed) through a number of discrete time steps according to a set of rules based on the states of neighboring cells. The rules are then applied iteratively for as many time steps as desired.

There are at least 21 implementations of SLEUTH in the United States and 11 in international regions. The model uses historical growth patterns derived from classified remote sensing data, like the National Land Cover Data (NLCD), and ancillary data to predict where land cover change, or urban growth, will occur in the future. Currently, we are focusing on the urban growth portion of the model. This has important implications for transportation planning as well as agriculture, wildlife, and water resources for decades to come. The modeling package has the ability to show what the landscape pattern will likely be at discrete periods decades from now based on past development patterns.

The pilot study area for this project is Lexington-Fayette County and portions of the adjacent counties of Clark, Bourbon, Scott, Woodford, Jessamine, and Madison. This multi-county region of Kentucky provides an opportunity for several reasons. This region has been experiencing land use conversion from forest/agriculture to urban uses. The region is experiencing an expected to continue increasing population and resulting housing unit capacity expansion. For example, Scott County is the third fastest growing county in the state in terms of population growth over the last six years (24%). Jessamine County population was up over 14-percent and Madison County at almost 11-percent making them the seventh and eleventh fastest growing counties in the state respectively. The Urban Service Area in Lexington-Fayette County has been in effect, with some modifications, for almost 50 years. This presents an unparalleled opportunity for predictive modeling since it is the oldest type of policy boundary in the nation. In addition, many entities are calling for regional cooperation and thinking relative to the economy and quality of life. Land use and transportation planning is central to quality of life issues.

Objectives of this project include: 
1.Utilize historical land use patterns to predict land use patterns considering topographic form and roadway infrastructure.
2. Create land use scenarios in an experimental environment depicting the next several decades.
3. Identify potential urbanization patterns useful for transportation planning activities in a multi-jurisdictional planning region.

Forest Fragmentation

Principal Investigator
Brian Lee, Department of Landscape Architecture

Project Summary
Identifying the Extent of Kentucky's Forest Fragmentation

Project Description
Forest fragmentation refers to the severance of tracts of forested land as a result of harvesting practices and clearing for agricultural lands, as well as the development of human habitat, roadway construction, and other human-influenced landscape modification such as resource extraction, and is an important consideration in a comprehensive landscape conservation portfolio. Habitat loss due to forest fragmentation is often cited as one of the most important reasons for the decline of biological diversity around the world as well as in Kentucky.

The original intent of the Forest Fragmentation project was twofold. The first objective was to develop a framework for automating a process to classify forest fragmentation and to quantify (size) interior forest blocks of greater then 1000 acres in Kentucky using the Environmental Systems Research Institute's (ESRI) ArcGIS ModelBuilder. The second objective was to apply that process to three classified datasets at two temporal points to characterize potential change: Kentucky GAP Analysis data (performed using Landsat Thematic Mapper (TM) imagery); 1992 National Land Cover Data; and 2001 National Land Cover Data.

The spatial approach used to conceptualize the model was similar to the way a sculptor might create a subtractive clay block sculpture. The form is a result of removing material from the whole, in this example, from forested lands. A moving window analysis was used to characterize the forested cells as patch, transitional, edge, or interior forest. Three maps (one for each dataset) were generated to illustrate the extent of forest fragmentation. Two charts illustrate the quantitative characteristics of the forests from the three datasets: amount of fragmentation by type (Patch, Transitional, Edge, and Interior), and detailed quantitative characteristics of Interior Forest. It is important to recognize that there is known error in each of the land cover/land use datasets according to the data originators; however it is believed that these issues would amount to small errors and not substantially change the results.

In 2008, these geospatial models were applied to the Kentucky Land Cover Change Detection 2001/2005 data to update Kentucky's forest characterization with the most recent classified data publicly available. A conditional statement was used to combine the 2005 data with the KYNLCD 2001 where change was not detected. As with the previous three datasets, large forest blocks (>1000 acres) were identified and forest fragmentation was classified, and in addition, forest blocks greater than 5,000 acres were queried as well as forest blocks greater than 20,000 acres were queried to show spatial distribution of the largest of the large forest blocks. When quantitatively comparing the 2005 forest data to the 2001 forest data, there were reductions in all of the basic descriptive statistics, and the forest fragmentation characterization showed an increase in the amount of patch and transitional forest while showing a decrease in edge and interior forest.

Determining Landscape Areas for Targeted Reforestation Efforts
In 2010, work is underway to utilize this forest block data in combination with other publicly available data in a modified ELESA approach for forest land (FELESA) to identify critical locations in Kentucky for focusing reforestation efforts. This project will use historical land cover data and other data to predict the probability of changes in the forested landscape for the next 100 years under a variety of scenarios. The study area chosen to depict potential alternative forest futures is the Eastern Kentucky Coal Field. This natural resource management project also has land-use planning implications.

The FELESA process evaluates land consistently based on stakeholder input to model guidelines. The system is flexible so that suitability analysis can be based on a variety of stakeholder objectives such as water quality protection, biomass/carbon accumulation, forest connectivity, interior forest expansion, or physical access to name a few. For a given land area, factors are rated, weighted, and combined, resulting in a single numeric reforestation suitability score. The score indicates the relative land value in relation to other forested and non- forested land for reforestation activities.

In a second suitability approach, another software application, Marxan, will be used to identify reforestation opportunity areas. Marxan is the most widely used landscape conservation planning software (Ball et al., 2009; Watts et al., 2009). It provides decision support to a range of conservation planning problems including developing multiple-use plans for natural resource management. Marxan is efficient and repeatable providing a number of options and encouraging stakeholder participation. These features provide users with decision support to achieve an efficient allocation of resources across a range of different uses.

Once the most suitable reforestation areas are identified, LANDIS II will be used to simulate forest change. LANDIS II is a spatially explicit landscape model designed to simulate forest landscape change over large spatial and temporal scales (Mladenoff et al., 1996; He et al., 1999). LANDIS has been used to simulate the dynamics of forest succession, seed dispersal, wind, fire, biological disturbance (insects and diseases), harvesting, and fuel management. Differing from most landscape models, LANDIS simulates multiple landscape processes in combination with the simulation of succession dynamics. The forest simulations will be quantitatively evaluated and subsequently evaluated by the experts involved with the reforestation suitability model.

This project will utilize and build upon precision resource management capacity for the state in four specific ways:

1 . Capitalize on the previous remote sensing and other geospatial data investments that are available from the Kentucky Geography Network.

2 . Apply land use suitability analysis relying on an Analytic Hierarchy Process in the context of a geospatial Delphi framework for identifying locations for the planting of 175 million trees in eastern Kentucky.

3 . Utilize the capabilities of one of the most widely used software applications that can identify potential conservation reserve locations (Marxan) to compare to the locations identified using a geospatial Delphi framework.

4 . Once suitable reforestation locations are identified, multiple simulated forest change scenarios over large spatial and temporal scales will be performed using LANDIS.

The Watershed Atlas Project

Principal Investigator
Brian Lee, Department of Landscape Architecture

Previous Assistant Researcher
Corey Wilson, Department of Landscape Architecture

Previous Undergraduate Researchers
Collin Linebach, Department of Landscape Architecture
Drew Heering, Department of Landscape Architecture

Project Summary
Landscape Characterization in Kentucky as an Indicator of Watershed Health

Project Description
Assessing the landscape from a watershed and waterway health perspective provides an indication of the quality of an area's land management decisions. The Watershed Atlas Project is an approach to waterway and watershed health assessment that utilizes existing publicly available geospatial data (Kentucky Geography Network and related sites) to visualize landscape indicator aspects of multiple watersheds simultaneously. Initial grounding for this work can be found in An Ecological Assessment of the United States Mid-Atlantic Region: A Landscape Atlas (1997) and has been modified and expanded for The Watershed Atlas Project. By looking at landscapes as indicators to infer watershed health, it is possible to modify land management decisions for the future to improve water quality and reduce the risk of future threats, and understanding how watersheds are statistically similar or different (cluster analysis) can aide in mobilizing a suite of protection or mitigation strategies.

The Watershed Atlas Project uses the Hydrologic Unit Code (HUC) as the fundamental unit of analysis. HUC-14s are the smallest unit of the United States Geological Survey's (USGS) hierarchical organization scheme for cataloging watersheds across the United States from the largest scale drainage basins in the nation, down to the smallest scale watersheds (The National Atlas, 2007). With this atlas, watersheds are characterized using data available from the Kentucky Geography Network in the Environmental Systems Research Institute's (ESRI) ArcGIS environment. ESRI's ModelBuilder was utilized with additional data (Clumpy Index) reported from Fragstats to measure the landscape according to more than 100 indicators (currently), organized by theme: Geographic, Geomorphic, Human, Vegetation, Riparian, and Specialty Indicators. The atlas provides a series of maps illustrating relationships among watersheds for each indicator as well as a detailed data table providing numeric values for each HUC and each indicator.

To date, The Watershed Atlas Project has been utilized to predict watershed health in the Licking and Kentucky River Basins, as well as for a seven Kentucky county region, including: Clark, Fayette, Jessamine, Madison, Owen, Scott, and Woodford counties; and an effort to assess the Licking River Basin. Historically, students in the capstone service-Learning Studio have applied elements of The Watershed Atlas Project to Boone, Campbell, Kenton, and Logan counties.

Watershed Prioritization
Next steps for the Watershed Atlas will work toward applying the data to on-the-ground approaches for protecting or improving watershed health. One approach to that process includes identifying priority watersheds in which to focus efforts. Watershed Atlas data are used to prioritize watersheds through a process which calculates a Z-score for several key landscape indicators which can be added and weighted based on areas of interest. A Riparian Agriculture scenario, for example, might weight the Riparian Agriculture indicator Z-score very high, at 85%, and include the Road/Stream Intersections, Stream Density, and Riparian Zone Slope indicators at 5% each to determine those watersheds which are high priority due to known watershed health issues related to waterway and riparian interactions with agricultural practices.

Water Quality Trading in the Kentucky River Basin
Aspects of the Watershed Atlas are being used to provide several types of geospatial analyses to a project by the University of Kentucky Department of Agicultural Engineering researching Water Quality Trading Opportunities for the Kentucky River Basin. The Watershed Atlas, in addition to two different versions of the Soil and Water Assessment Tool (SWAT), is being utilized to indicate where Nitrogen and Phosphorus emission credit is needed and where the credit potentially exists in the river basin. It is important to understand how the two nutrients behave in the Kentucky River Basin so that a more realistic spatial trading market could be explored for feasibility. For Landscape Architecture, this project is particularly important because it reinforces the linkages between the urbanized and rural areas of the state while integrating natural and social systems that have implications for how thousands of acres of the state might look one day. 

Housing Price Study

Principal Investigator
Brian Lee, Department of Landscape Architecture

Undergraduate Researcher
Drew Heering, Department of Landscape Architecture

Project Summary
A study of the effects of the Lexington-Fayette County Urban Service Area on housing price.

Project Description
This is an investigation into the housing price relationships inside and outside the Urban Service Area (USA) of Lexington-Fayette County, Kentucky and adjacent counties. The Lexington-Fayette County USA was instituted in 1958 and is generally recognized as the oldest in the country. This study replicates the hedonic price framework approach of a Portland, Oregon study for Lexington-Fayette County using census block group data from 2000. What are the major variables that will largely determine housing prince in the seven county region? Does the USA boundary influence housing price or are there other variables? A paper for this project is in the later stages of development and is expected to be ready for peer review in the coming months.

Metropolitan Sprawl Gradient

Principal Investigator
Brian Lee, Department of Landscape Architecture

Undergraduate Researcher
Drew Heering, Department of Landscape Architecture

Project Summary
An approach to measuring urban sprawl by proximity, and housing unit and population density.

Project Description
The Metropolitan Sprawl Gradient (MSG) approach tested in the Lexington-Fayette County region of Kentucky measures sprawl by combining measure of proximity with housing unit and population density. Urban sprawl has been characterized by the proximity to a variety of public use facilities such as the post office, supermarket, library, etc., as well as the density of people and/or housing units. This project uses proximity and relative densities to help visualize an urban development surface. The proximity to public services is determined using ESRI's ArcGIS tool, Euclidean Distance, from point, line (bus route), or polygon (parks) features representing those services. The study utilized publicly available data from the Kentucky Geography Network, along with data created using heads-up digitizing in conjunction with Google Maps and Local Live Maps to geolocate places of interest.

Walkability and Bikeability

Principal Investigator
Brian Lee, Department of Landscape Architecture

Undergraduate Researcher
Jared A. Cunningham, Department of Landscape Architecture

Project Summary
Walkability and Bikability Assessment and Least Cost Improvement Methodology

Project Description
The Centers for Disease Control and Prevention and the Dietary Guidelines for Americans 2005 recommend that children and adolescents frequently participate in at least 60 minutes of moderate-intensity physical activity, preferably on a daily basis. One way to increase physical activity is to incorporate it into the child's daily school commute. However, neighborhoods have often been designed with the automobile exclusively in mind. Consequently, children walking or bicycling to school is not always a safe alternative to the car or school bus. With ArcGIS and the Spatial Analyst extension, students identified dangerous walking and bicycling areas, proposed design safety solutions, and evaluated alternatives for improving adverse conditions.

The study area covered approximately .60 square miles and contained neighborhoods built from the early 1900s through the 1960s. The destination within the study area was an elementary school located near the center. The student team began by developing a neighborhood audit for streets and intersections that used the Likert Scale rating system. A Likert Scale is a qualitative rating system that is often used to measure judgments. Ratings are indicated along a continuum, in this case on a scale of 1 through 5, indicating the quality of the street or intersection in terms of safety or ease of movement.

Basemaps were prepared from publicly available geospatial data from the Kentucky Geography Network to aid in field data collection and subsequent analyses, and the ordinally scaled qualitative audit data was entered into ArcGIS as attributes for a shapefile. The shapefile was used for initial audit visualization and analysis. Using the audit data, the team was able to produce a series of maps depicting current walking and bicycling conditions within the project area and another series of maps that included straight-line walking distance, topographic slope, and children density.

The initial audit analyses and ancillary data helped visualize where problems existed and allowed the team to debate strategies on how to improve the walkability and bikeability for the neighborhood school. To determine the relative costs of going to and from school for the children, the team used the Cost Distance tool of the Spatial Analyst toolbox in ArcView 9.1 with the ArcGIS Spatial Analyst extension.

The team evaluated three strategies to improve conditions based on a combination of the slope, audit, and children density. The final analyses evaluated three solutions indicating a 12 percent, 48 percent, and 50 percent increase in overall study area walkability and bikeability. Because 50 percent of the evaluation was based on land slope, none of the three proposals for existing neighborhoods improved walkability and bikeability 100 percent. The student team was able to propose strategies that included a range of on-the-ground changes. Potential changes included adding proper curb cuts, painting crosswalks, adding stop signs, modifying vegetation, and adding sidewalks. These modifications ranged in cost and difficulty of implementation. The students also evaluated not only the types of strategies but also where to implement them.

This analysis and proposal evaluation method allowed for more informed debate among the team members on the best ways to improve walkability/bikeability conditions. Simply by changing the cost friction surface and rerunning the Cost Distance tool, the students could quickly test a variety of design solutions. Although not specifically part of this academic exercise, a cost-benefit analysis could be included for additional help in making decisions. This method could be easily adopted for use in projects underway or planned by government and nonprofit organizations as well as the private sector.

Enhanced Landscape Evaluation and Site Assessment

Principal Investigator
Brian Lee, Department of Landscape Architecture

Undergraduate Researcher
Collin Linebach, Department of Landscape Architecture

Summary
An established framework for making land-use decisions was implemented in ModelBuilder to provide a more holistic approach to community planning.

Project Narrative
Land-use decision making is an increasingly important capability for many local governments. The Land Evaluation and Site Assessment (LESA) framework, developed by the Natural Resources Conservation Service (NRCS), has been used by hundreds of communities across the country to help make decisions, particularly in the context of farmland protection programs. Land evaluation (LE) and site assessment (SA) are the two aspects of this approach. The LESA framework aids land-use decision makers by rating lands based on land evaluation and site assessment rating factors. The factor rating system is determined by a committee composed of local people who are familiar with the landscape and analysis objectives. A key benefit of using LESA is that it provides a clear and consistent decision tree-type framework for land rating.

LE primarily rates soils. For this Central Kentucky study region, most of the data was obtained from the Kentucky Geography Network, and soil data came from the NRCS. The best soils were given a high ordinal rating (10), while less agriculturally rated soils were assigned a lower rating (1). The SA section is often more complex and can consist of three general categories: nonsoil agricultural factors, development pressures (e.g., road and sewer proximity), and other public values (e.g., historic, scenic). Like the LE portion, each SA factor described in this article was rated relative to agricultural generalization to demonstrate the functionality of this approach.

In this Enhanced LESA (ELESA) implementation, The Environmental Systems Research Institute's (ESRI) ModelBuilder was used to automate the rating as well as factor weighting as described in the LESA guidebook by Pease and Coughlin. The model can be thought of in terms of three primary components: Land Evaluation (LE) factors, Site Assesment (SA) factors, and overall factor weighting. Raster-based ELESA allows the landscape to be thought of as a continuous surface of LESA process rating. This rating visualization can help identify suitable acreage for people interested in agricultural land for a variety of reasons. The factor rating and weighting can be tailored to meet the desires of an individual or an entire community. The introduction of ModelBuilder benefits the process by allowing different rating and weighting combinations to be explored during public meetings depending on land-use problems. The model has the ability to include land protected by land trusts or other entities under agreements such as conservation easements and purchase of development rights programs. The model can also visualize how ELESA ratings change relative to agricultural uses when urbanization proposals are anticipated.

GIS has enhanced the power of LESA for landscape analysis and participatory planning/decision making. The regional/landscape scale analysis performed in this ELESA model allows for broad scale and flexible analysis. ELESA has the potential for areawide planning applications, and this model could be adapted for watershed-based applications and modified for use with ArcGIS Server.

Stream Visual Assessment Protocol

Principal Investigators
Brian Lee, Department of Landscape Architecture 
Christopher Barton, Department of Forestry and Natural Resources
Barry W. Kew, Department of Landscape Architecture (now at University of Cincinnati)

Assistant Researcher
Corey Wilson, Department of Landscape Architecture,

Undergraduate Researcher
Ritchie Katko, Department of Landscape Architecture
Collin Linebach, Department of Landscape Architecture
Drew Heering, Department of Landscape Architecture

Project Summary
Watershed Assessment and Method Developement

Project Description
Evaluating even the basic condition of aquatic ecosystems across the Commonwealth is a financial and time consuming endeavor. Furthermore, detailed biological and chemical sampling of stream conditions is not always possible because of limited financial, technical, and personnel resources. However, a relatively quick method of performing a first pass stream assessment based solely on visual inspection of quality has recently been developed and published by the Natural Resources Conservation Service (NRCS) called the Stream Visual Assessment Protocol (SVAP).

The number of watershed councils is increasing in Kentucky. Typically these organizations do not have plentiful financial or technical resources. An early challenge for these organizations is performing an assessment of watershed conditions. This goal of this project was to develop a SVAP geospatial data collection application for ESRI's ArcPad mobile GIS for use in basic watershed assessments across the Commonwealth that is relatively inexpensive to deploy and use by these watershed councils.

This pilot project, located at Robinson Forest in Perry, Breathitt, and Knott counties, will measure geomorphic descriptive characteristics for the Robinson Forest watersheds for sample locations both pre- and post- timber harvest. The intent is to develop a method utilizing geospatial technology and field collected data capable of predicting aquatic conditions as assessed by SVAP from watershed geomorphic and land cover characteristics. The development of this method of predicting aquatic ecosystem condition has the potential to be useful in watershed assessment and education efforts in other parts of the Commonwealth as well as other Appalachian areas. Robinson Forest was an ideal place for this research because it complements ongoing research projects related to stream side management zones.

Gap Change

Principal Investigator

Brian Lee, Department of Landscape Architecture
M. Keith Wethington
Kentucky Deptartment of Fish & Wildlife Resources

Undergraduate Researcher
Collin Linebach, Department of Landscape Architecture

Project Summary
Terrestrial Vertebrate Habitat Change Characterization in Kentucky, Pilot Project

Project Description
Habitat loss and fragmentation is often cited as one of the most important reasons for the decline of biological diversity around the world as well as in Kentucky. Understanding how the landscape is changing is essential for identifying viable habitat for wildlife management. This project leveraged the existing methods and data of the original Kentucky GAP Analysis (Wethington et al., 2003) sponsored by the United States Geological Survey and principally performed by Kentucky Department of Fish and Wildlife Resources and Murray State University researchers.

Gap analysis is a scientific method for identifying the degree to which animal species and natural communities are represented in the present-day mix of conservation lands. Those species and communities not adequately represented in the existing network of conservation lands constitute conservation "gaps." The purpose of the Gap Analysis Program (GAP) was to collect broad geographic information on the status of species and their habitats in order to provide land managers, planners, scientists, and policy makers with the information they need to make better-informed decisions (Gap Analysis Program, 2000).

The original Kentucky GAP Analysis was performed using Landsat Thematic Mapper (TM) imagery that is almost two decades old. Subsequently, habitat suitability maps currently used are also more than a decade old. There is a need to update the basic maps for management and planning purposes. In addition, there is currently not the capability to more holistically understand how the landscape is changing over time in relation to wildlife habitat across the state particularly for species of special concern.

This pilot project utilized landcover data available from the Kentucky Landscape Snapshot project based on 2001 and 2005 Landsat imagery. These new landcover data provided the opportunity to update wildlife habitat suitability maps. The 2001 and 2005 Environmental Monitoring and Assessment Program's (EMAP) Hexagons were also used to quantify the amount and spatial configuration of terrestrial vertebrate habitat. This provided the opportunity to test for differences in amount, location, and spatial configuration of suitable wildlife habitats and to grossly characterize how terrestrial vertebrate species habitat is changing across the state.

Objectives
1. Produce wildlife models based on 2001 and 2005 landcover data. 
2. Quantify amount and spatial configuration by EMAP Hexagons 2001 and 2005 of terrestrial vertebrate habitat.
3. Test for quantitative differences in habitat metrics of Objective Three.
4. Visualize advancing and declining habitat via EMAP Hexagons.