CICEET Progress Report for the period 9/02/07 Through 3/01/08

Project Title: Integrating Geospatial and Web-Based Technologies to Improve Land Use Planning in Coastal New Hampshire
Principal Investigator(s): Fay Rubin, Complex Systems Research Center, UNH
Project Start Date: September 1, 2007

Figures

Figure 1

Figure 2

Figure 3

Within this broad framework, our specific project objectives for this period were as follows:

• Assemble existing geospatial data sets and develop required data elements to support remaining project objectives
• Initiate refinement and implementation of SUBNET model
• Initiate identification and adaption of resource models to ESRI ArcGIS ModelBuilder environment
• Initiate construction of web tools to display and query data

2. Initiate refinement and implementation of SUBNET model
This project objective requires the evaluation, refinement, and implementation of the Simulating Urban growth with Bid-rent and NEural network Tools (SUBNET) land use change forecast model as a decision support tool for land use planning. The two major components of the model predict the magnitude of conversion to developed land (demand) and the likely location of the developed land.

3. Initiate identification and adaption of resource models to ESRI ModelBuilder environment
A number of coastal communities have access to ArcGIS, and will be eager for land use planning tools that can be easily incorporated into their existing technology infrastructure. While many GIS techniques have been developed to map and analyze community resources, they have not been applied widely due to the complexity of the series of steps needed to complete the analyses. Instead of forcing users to go through an intricate series of tool activations and adjustment of analysis parameters, ModelBuilder provides a means for distributing an entire sequence of steps as a ready-made tool. Thus, our objective is to identify the existing resource models that will most benefit the coastal community, and adapt those models to the ModelBuilder environment.

4. Initiate construction of web tools to display and query data
The design of web tools for displaying maps of the model output (development pressures), threats to water quality and land use information, constituted the 4th major objective of the project.

Progress on each objective is reported below.

Tasks to meet objectives
Objective 1:
The first substantive task was to complete the development of detailed land use data for Rockingham County, in southeast New Hampshire. High resolution (1-foot), color imagery, acquired in the spring of 2005, was available to serve as the source for the data development. Using standards and guidelines produced through a coordinated effort of the regional planning agencies in NH, the NH Office of Energy and Planning, and CSRC staff, land use data is being interpreted from these source images by CSRC and RPC.

CSRC’s second substantive task was to begin development of a detailed impervious surface data for Rockingham County, using object based image segmentation processing software (eCognition) applied to the same source imagery.

Thirdly, Applied GeoSolutions (AGS) was tasked with acquiring and processing the road network data, as well as assembling other data sets required for their modeling activities.

Objective 2:
In addition to the data acquisition tasks noted above, AGS planned to review and evaluate the literature to determine how accessibility to markets is handled by other land-use/land-cover change models during this performance period.

Objective 3:
The first task to meet the objective was to identify GIS analyses useful for coastal land use planning based on a combination of research and input from stakeholder discussions. Once the universe of possibilities is refined to a targeted subset of useful models and analyses, the second major task is to code these analytical techniques using the ESRI ModelBuilder format, and make the resulting code available to communities.

Objective 4:
Tasks to meet the objective included the construction of web tools to display and query data, and the development of three-dimensional data visualization techniques.

Progress on Tasks
Objective 1:
Land use automation for Rockingham County is underway. Currently, twenty eight of the thirty seven towns in the county have been automated and incorporated into a seamless data set (see Figure 1). Two of the remaining nine towns are in final QC and therefore near completion, with mapping in progress for the balance of the county. Basic interpretation and mapping is being completed by regional planning commission partners and CSRC staff, with CSRC reviewing and performing final QC for the entire area. The final land use data layer comprises 55 detailed use categories, as illustrated in Figure 2.

Figure 1. Status of Land Use Automation for Rockingham County, NH.
Figure 2. Detailed Land Use Mapping for Hampton Falls, Rockingham County, NH.

We have not initiated the development of the impervious surface data set.

AGS staff researched and reviewed numerous roads datasets and determined that the best data for the project is the TeleAtlas streets network provided by ESRI (circa 2002; based on 2000 era US Census Bureau TIGER line data). These data include speed limits for all roads in the US and are thus consistent across the entire study area (job markets from Boston, MA to Manchester, NH, to Portland, ME).

In addition to the data sets described above, AGS staff has acquired the following data sets for the study area:

• Roads (ESRI/TeleAtlas street network)
• Wetlands (National Wetlands Inventory)
• Topography (USGS National Elevation Dataset)
• Soils (NRCS)
• Conservation and Public Lands (TNC Secured Areas)
• Land Use (GRANIT ­ 1962, 1974, 1998)
• Real estate transactions data

Objective 2:
Staff at AGS has made significant progress towards refining the SUBNET land-use prediction model. Thus far, efforts have focused on the location side of the model, and in particular, on a refined approach to modeling access to job markets.

To determine how accessibility to markets is handled within the land-use / land-cover change community, AGS evaluated numerous land cover / land use change models. These models fall into two major categories with respect to accessibility:

• Purely spatial / physical models (i.e. accessibility to markets is not considered explicitly and predicted change is based solely on physical factors such as slope, soil type, or proximity to existing development; examples include SLEUTH, GEOMOD, and the Land Transformation Model).
• Economic models (i.e. predicted change is based largely on economic input-output models; examples include UrbanSim, TRANUS, and METROPILUS)

As a combination of these two approaches, the refined version of SUBNET (as with the original SUBNET model) will approach the question of the attractiveness of a locale to be developed based on the bid-rent theory of economic geography; i.e. locales closer to urban job markets will be more likely to be developed.

Accessibility will be handled from the point of view of the worker. For each locale in Rockingham County, for any drive time increment (based on our road network dataset), we will calculate the number of accessible jobs (based on US Census tract data).

An example of this approach to accessibility is illustrated in Figure 3. For this tract (in Exeter, NH) one can see that the majority of accessible jobs lie between 35 and 75 minutes drive time. The small peak at 35-40 minutes (~176,000 jobs) represents access to the Manchester, NH and the Lawrence / Haverhill business districts. The peak at 55-60 minutes (~640,000 jobs) represents access to the Nashua, NH and Boston, MA business districts as well as the “office park” suburbs northwest of Boston (such as Woburn, Burlington, and Bedford, MA).

Figure 3. Accessibility to job markets from Exeter, Rockingham County, NH.

Objective 3:
Shane Bradt, UNH Cooperative Extension, has initiated the compilation of GIS land use planning tools that currently exist through conversations with various national organizations related to planning and/or geospatial tool implementation (NOAA Coastal Services Center, Nonpoint Education for Municipal Officials, National Association of Counties, National Geospatial Technology Extension Network). Shane will be meeting both with the town of Exeter (our municipal partner) and with the Rockingham Regional Planning Commission in the next month to discuss their current use of GIS and evaluate their needs for GIS tool development. The results of all of these conservations will be used as a starting point in the stakeholder forums on land use planning tools which will be held this summer.

Shane has attended training sessions in several of the tools already identified as potentially useful (Community Viz, Habitat Priority Planner, Impervious Surface Analysis Tool, Landscape Fragmentation Tool, Nonpoint Source Pollution and Erosion Comparison Tool) and is working toward the development of regional coefficients for several tools (Impervious Surface Analysis Tool, Nonpoint Source Pollution and Erosion Comparison Tool). Shane is also working to incorporate several of these tools into current or new training workshops as part of the UNH Geospatial Technologies Training Center (Community Viz, Habitat Priority Planner, Landscape Fragmentation Tool, ModelBuilder). In addition to providing a more diverse range of GIS workshop topics to our traditional clientele, the inclusion of these tools will provide a solid basis for the eventual development of targeted trainings for land use planners in the second year of this project.

Objective 4:
AGS has started development of a prototype website capable of dynamically displaying model output. Currently, the site is constructed using a combination of Open Source tools including OpenLayers, GoogleMaps, PostgreSQL, and some html/javascript. The building blocks can be seen at http://subnet.homedns.org (best if viewed in Firefox). The goal of this initial website is to solicit feedback from end users on the tool before the full prototype is constructed.

No activity was anticipated during this project period on the data visualization task.

Difficulties
Objective 1:
Development of the impervious surface data has been delayed. Because the project start date was later than expected, training of CSRC staff in using the targeted software (eCognition) will not begin until late summer, 2008. Despite this delay, we anticipate delivering the impervious surface data in sufficient time to contribute to the remaining project objectives.

There have been no unexpected difficulties in accomplishing the remainder of the tasks.

Objective 2:
There have been no unexpected difficulties in accomplishing the scheduled tasks.

Objective 3:
There have been no unexpected difficulties in accomplishing the scheduled tasks.

Objective 4:
There have been no unexpected difficulties in accomplishing the scheduled tasks.

Data Generated to date
Several geospatial data sets are under development. Once finalized, those that are not restricted will be documented and posted on the GRANIT web site (www.granit.unh.edu) for public access.

Project Objectives for Next Reporting Period

Objectives
1. Completing the detailed land use data set for Rockingham County
2. Initiating development of the impervious surface data set
3. Continuing to refine and implement the SUBNET model
4. Prioritizing tools to adapt to ModelBuilder format
5. Continuing the development of web tools

Work Plan to Meet Objectives
Objective 1:
Continue close coordination between CSRC staff and Rockingham Planning Commission staff in developing the land use data set. As in the past, CSRC will receive all preliminary data and will perform final QC.

Objective 2:
Participate in eCognition training in Morristown, NJ, in August of 2008. Subsequent to this formal training, CSRC will begin development of the impervious surface data.

Objective 3:
Complete the fully defined road network, which will require refining impedance turns (left/right/U-turns) and traffic volume, and developing 1962 and 1974 road network data sets based on historic land use data.

Refine accessibility to job markets dataset by testing drive time increments to determine ideal time resolution; refining the spatial unit at which to determine accessibility (e.g. census tract, census block, etc.); testing the theoretical framework (i.e. bid-rent) with the new approach to accessibility against real-estate transactions and other factors influencing the cost of land; and developing an “accessibility surface” to be used in the ANN where each band represents the number of accessible jobs within the given drive time increment.

Refine demand component of SUBNET model through the acquisition of the NH employment data and REMI economic forecast models

Develop regression equations to predict built area based on current employment data; develop predictive model based on equations and REMI economic forecasts.

Objective 4:
Meet with partners from Town of Exeter and Rockingham Planning Commission to evaluate needs for GIS tool development

Host a forum on land use planning tools for coastal planning community

Objective 5:
Continue web tool development by bringing relevant data layers into the prototype (land cover, appropriate vector data, political boundaries, etc.) and connecting a scaled-down version of the SUBNET model to the website, thereby enabling users to modify select parameters and explore ramifications of alternate planning decisions.

Expenditures
Expenditures have been in the range anticipated during the first project period.

What Else?
The overall project design relies heavily on the input of our local project partners ­ primarily the Town of Exeter and the Rockingham Planning Commission. However, we have realized that their input will be most valuable in later phases of our work as our project concept becomes more refined. In the land use modeling area, for example, local input will be invaluable in determining which elements of the modeling will be most useful as web-enabled components, as well as in looking at the local impacts of the forecasts. In the identification of resource models for adaptation in ModelBuilder, local feedback will help us to evaluate the utility of various approaches once the project team has identified the existing models, defined their data input requirements, and performed an initial assessment of their feasibility.