Habitat modeling based on the USDA Forest Service's Forest Inventory and Analysis Program

(the document is a compilation and summary of the proposal submitted to National Council for Air and Stream improvement, Inc. and USDA Forest Service in 2002)

Introduction and Background

The ongoing Forest Inventory and Analysis Program collects and processes along with timber-related information many environmental variables, so not only forest products industries, but also a wide spectrum of wildlife and conservation-oriented users can rely on the information to be inferred from it.

To answer questions about natural resources sustainability, availability, and development, one needs to conduct complex analysis involving proper modeling of changes over time. This includes using explicit assumptions concerning forest regeneration dynamics and management activities, clear assumptions regarding future land use changes, and also by taking habitat required by a variety of important wildlife species into account.

Through the proposed project we want to provide a scientific and operational basis for realistic analyses of wildlife habitat. We want to develop new approaches for using data provided by the Forest Inventory and Analysis Program, in conjunction with various additional sources, to assess wildlife habitat and for analysis of effects of various natural resource development scenarios on spatial aspects of habitat requirements for game and non-game species.

Objectives

We propose this study as a pilot project with the primary objective of investigating methods of extracting information on spatial aspects of habitat availability for game and non-game species based on FIA data analysis. To control costs and still produce useable results within the two-year timeframe of the project, we propose to use the FIA data for the State of Georgia in a case study style of approach. Associated objectives are to develop a methodology for using hierarchical data from the FIA program, in conjunction with other available data and remote-sensed sources of information, suitable to assess wildlife habitats on local, regional and national scale.

We are going to develop methods for extracting information from FIA data that pertains to spatial aspects of habitat requirement for various species. We propose to use a spatially explicit, landscape-scale simulation model, fed by the FIA data populated on the processed satellite (Landsat TM) images and other available GIS data. Spatial capability of the model wide used in conjunction with the available data sources allow us to include not only forest inventory related data, but also attributes of the physical environment such, as soils, topography, water resources, etc. Simulations of different scenarios involving different targets and types of disturbance will be used to simulate the dynamics of changing habitat requirements for game and non-game species over time.

We want to develop a methodology for analysis of effects of various assumptions and environmental changes on changes in wildlife habitat. This would include testing of impacts of riparian zones and road buffers, as well as various land use changes, on wildlife habitat availability. The analysis will include consideration of necessary activities to provide certain wildlife habitat on selected areas. Developed procedures will include linkages among ecological data, analytical procedures, and landscape characteristics that can be determined from FIA data. Relevant additional ecological data will come not only from existing free or reasonably priced sources, such as EPA (NLDC, C-Cap, NALC), USGS (e.g., STATSGO), EMAP, and other databases, but also from available remotely sensed information (satellite imagery) processed with geostatistical methods. Using these techniques will assure that operational implementation of habitat models for use by FIA and its partners is possible on local, state-wide, regional and national bases.

Description of the proposed study plan

We will use the spatially explicit forest estate model OPTIONS to forecast polygon-level characteristics of the identified cover types on an annual basis. Temporal changes in each polygon and sub-polygon will be estimated by simulation of scenarios representing interpretations of forest management practices and conservation initiatives. This simulation capability enables the results of simulated scenarios to be used to forecast present and expected future habitat conditions and spatial habitat distributions for a wide variety of wildlife resources in considerable detail. Because OPTIONS dynamically maintains multiple GIS layers with each layer having its own management criteria and management objectives, individual GIS layers can be used to establish criteria and objectives for a wide variety of spatially sensitive land management objectives and resources, such as for individual wildlife species, special management zones (such as riparian areas), aesthetics, steep slope protection, wetlands protection, etc. The multi-layer GIS capability allows, through the use of simulations, to dynamically and simultaneously evaluate data, parameters, polygons and land management strategies for many uses.

While the FIA provides data at the state, region, or national level, all other levels of resolution can be actively reviewed and analyzed. Conservation value criteria, such as terrestrial wildlife habitat, riparian and visual/aesthetics zones, soils/unstable slope protection, featured/listed wildlife species and rare plant habitat protection, rural/urban zones, maintenance of a specified distribution of seral stages across forested landscapes, specialized stand structures and/or stand conditions related to all of the above criteria, both in terms of abundance and distribution, maintenance of existing gene pools with respect to un-disturbed, representative, older, mature forest types and minimally-disturbed forested ecosystems can be actively addressed in the projects.

In managing and evaluating lands for a broad range of conservation values, two broad tasks are involved. First, it is necessary to identify the conservation values present on a landbase, that come from a wide range of interest and resource management groups, to define and locate areas of interest on the basis of various criteria. The list of parameters used to define or identify lands which fall under one or more of the above criteria may include parameters used to describe forest cover, topography, soils, stream class, location and spatial juxtaposition. In other words, many different parameters and types of measurement are used to identify areas of interest. Second, it is necessary to establish desired benchmarks relative to the abundance and distribution of the conservation values. It is also necessary to be able to forecast the effects of land management alternatives on both the abundance and distribution of all conservation values, so that the results of the alternatives can be compared to the benchmarks.

When it comes to operationally managing a landbase, the list of parameters that can be used to manage and monitor environmental performance is much more limited than the lists often used to define conservation values. Parameters must be quantitatively measurable and sampled across the entire landbase using a statistically rigorous sampling system in order to be applicable to land management. Then, a methodology must exist for the development of a baseline of performance and to enable future forecasts of parameters to be made as a function of the management planning process. Finally, the baseline and forecasts must be monitorable. All of these characteristics are necessary in order to develop land management plans against which future performance can be measured.

For these reasons, the parameters used to measure and monitor performance for the forested components of habitats is based upon common forest inventory parameters. The measures for other physical resources present (such as water resources) are based upon comparative measures such as, in the case of water, streamflow, water quality (many measures), bank stability, headwaters slope stability, and characteristics of pools, riffles, gradient classes, stream's bottom composition, etc. The key parameters for these resources are linked to key forest inventory parameters and both the values of parameters for other resources as well as the forest inventory parameters are monitored over time for the identified portion of the area, which affects these other resources. The FIA data should be able to provide parameters meeting many of these criteria.

In addition, numerous data sources exist that can be used to link wildlife species to specific habitat characteristics that relate to FIA parameters. For example, the Georgia Department of Natural Resources (DNR) and the Chattahoochie-Oconee National Forest both maintain databases of point count surveys for birds. These consist of both bird abundance data and habitat characteristics that either are the same or similar to FIA parameters. We are very familiar with these databases because we set up one (Georgia DNR) and previously analyzed data from the other. Although the number of species used would have to be limited, priority species from assessments such as those made by Partners in Flight would be selected for analysis on a priority basis. Also, suggested indicator species that serve as indicators of ecosystem health for the ecosystems in which they live should be investigated. For example, the Louisiana waterthrush, as well as various amphibians, has been suggested as an indicator of riparian ecosystem health. For virtually every ecosystem, there is at least one bird species that is similarly tied to that system. Both the Partners in Flight Planning Process and the use of indicators of ecosystem health represent ecosystem approaches to management. These would be priority species for our analyses. Economically important species such as game species also would be investigated.

In some scenarios, we will have to make literature-based assumptions about the effects of various disturbances on wildlife. For example, the effects of roads or other "hard edges" on various wildlife species have now been well studied. From the literature, a range of effects at smaller scales can be hypothesized and tested over larger scales using our methodology. Disturbances that can be investigated include various forest management options, urbanization, prescribed burning options, and effects of natural disturbances.

We have already started a broad-based research analysis of long-term sustainability of forest resources in Georgia. Thanks to the capability of using the simulation model, OPTIONS, with spatially populated FIA data, utilizing existing sources of economic and environmental information, we are able to assess impact of different scenarios on timber supply not only on large scales (state-wide or national), but all other scales of interest, such as county or even single stands/polygons. Main information about forest stands and other environmental variables come from the processed FIA database. Existing information is distributed spatially using ancillary data that are available free of charge from different agencies (as e.g., EPA, USGS, USDA Forest Service, Bureau of Census, etc.), mostly through their Internet servers, or can become available from other related programs or purchased with reasonable investment (e.g., Landsat TM imagery for the area of interest).