A GIS TOOL FOR EVALUATING THE IMPACT OF PROPOSED CUTS ON RED-COCKADED WOODPECKER HABITAT

Transcription

1 A GIS TOOL FOR EVALUATING THE IMPACT OF PROPOSED CUTS ON RED-COCKADED WOODPECKER HABITAT Donald J. Lipscomb, Research Specialist Department of Forestry and Natural Resources, Clemson University Clemson, SC Thomas M. Williams, Baruch Institute of Coastal Ecology and Forest Science Clemson University, PO Box 596, Georgetown, SC, ABSTRACT The 2003 Red-cockaded Woodpecker (RCW) Recovery Plan changed the standards for good quality foraging habitat significantly. Instead of the 5 or 6 criteria previously used, the new guidelines use 11 to 16 to evaluate foraging habitat. An early attempt (2003) to evaluate RCW foraging habitat using 11 criteria on partitions mapped using RCWFAT resulted in 97 percent failure of RCW habitat to meet the guideline for both increasing and declining populations. The realization that the requirements for good quality foraging habitat in the plan were really ideal goals resulted in attempts to develop a score or foraging index placing habitat into categories ranging from poor to excellent. At the time of this abstract the aforementioned scoring system is still being developed by the U.S. Fish and Wildlife Service. In this paper we present a GIS program and a proposed index system built on data that can be derived from forest inventory, but with parameters chosen that correspond to 12 of the core criteria in the 2003 recovery plan. The indexing system proposed appears sufficiently sensitive to be used at any level from the basic management unit to a whole forest. The GIS model contains the tools previously in the RCWFAT AML program plus added interfaces to allow evaluating the impact of different cutting intensities on RCW foraging habitat. Results are stored in the partition attribute tables and can also be reported (output) as tables and text files. The interface for evaluating cutting intensities is interactive at the stand or basic unit level, however it will accumulate scores and evaluate the impact for any selected set of stands in a data set. The process performs a set of evaluations for before and after purposed stand alterations evaluating the change in habitat quality on the fly and updating attribute tables for the exported data sets. KEYWORDS. Red-cockaded woodpecker, recovery plan, foraging habitat, impact evaluation. INTRODUCTION The Second Revision Red-cockaded Woodpecker (Picoides borealis) Recovery Plan (USFWS, 2003) specifies criteria to evaluate foraging habitat that are considerably changed from the In Prisley, S., P. Bettinger, I-K. Hung, and J. Kushla, eds Proceedings of the 5 th Southern Forestry and Natural Resources GIS Conference, June 12-14, 2006, Asheville, NC. Warnell School of Forestry and Natural Resources, University of Georgia, Athens, GA. 139

2 former guidelines for management of habitat (Henry, 1989; USFWS, 1985). These revisions are based on findings by Beyer et al. (1996) and Wigley et al. (1999) that the main foraging criterion (8490 ft 2 basal area of pines > 10 DBH, within ½ mile of the cluster center) was not related to measures of woodpecker success. The new criteria are aimed at creating forest structure that benefits the woodpecker (James et al., 2001). These changes will impact approximately 1.6 million acres of federal lands to meet the recovery population goals. In 1998 we began development of RCWFAT (Red Cockaded Woodpecker Forage Analysis Tool) and used it to evaluate habitat on a variety of populations throughout the Southeast (Lipscomb and Williams, 2005). It has proved most useful for large dense populations found on Department of Defense (DOD) installations. Prior to 2003 it used the Henry (1989) guidelines to examine habitat quality of RCW clusters. Since 2003 we have modified the program to include the criteria established in the Recovery Plan, page These modifications were tested on data sets from the Oakmulgee Ranger District of the Talladega National Forest and Fort Bragg. These two sets reflect a healthy growing population on Ft. Bragg and a declining population on the other, requiring major efforts in habitat restoration. Despite the differences in population status the percentage of clusters meeting the criteria on the Oakmulgee (2.8%) was not much different from Ft. Bragg (3.8%) (Lipscomb and Williams, in press). These results are indicative that the habitat criteria in the Recovery Plan describe ideal habitat. Since the Recovery Plan presents only the ideal habitat for RCW foraging, few forests meet this high standard. Even the excellent habitat of Ft. Bragg did not meet this standard. In fact, only 18.8% of all stands (in RCW foraging areas) on Ft. Bragg met all of the criteria. Most of the 81.2% that did not, failed to meet two or more criteria. It is unlikely that any single silvicultural treatment would improve these dismal results. Additionally, the low values of success could easily obscure habitat decline. During 2004 a group of stakeholders at Ft. Bragg have been striving to develop a method to rank habitat that does not meet the goal of the criteria in the Recovery Plan. The basic plan has been to create a range of values for each criterion that represent five categories from very poor to meeting the Recovery Plan goal. Table 1 is similar to one of these plans with score of 1-5 for each criterion and the score corresponding to a range of values of the criterion that represent the gradient to poorer habitat. In addition there were similar tables used to score the condition of foraging areas for a single RCW cluster. We (Lipscomb and Williams, in press) used three of these proposed systems on data from both Ft. Bragg and the Oakmulgee. We found that that stand scoring was relatively robust and did not differ between alternatives. However, the efforts to combine the stand scores to evaluate clusters were undermined by the complexity of interactions between multiple stands and subtle differences in technique. Small changes in cluster criteria changed results much more than stand scores and produced highly inconsistent results. We have been led to propose a much simpler technique to evaluate impacts of management (cutting) on RCW foraging habitat. METHODS Our goal in this paper is to propose a simple method to produce a numerical value that represents the value of forest areas as RCW foraging habitat. The value we are calling Equivalent Good 140

3 Quality Foraging Habitat (EGQFH) is a product of the score of a stand and its area divided by the score of a stand that meets all Recovery Plan criteria. EGQFH = (stand score x acres)/ maximum stand score (with units of acres) The idea behind this index arises from a common occurrence, documented by Wood et al. (1985), that home ranges of RCW expand as the quality of forage declines. Since the index units are acres it can be accumulated for any geographic area the user chooses. The technique we are proposing was developed into a tool for ArcGIS 8.3 (ArcMap) and is added to a toolbar that has the other RCWFAT tools on it. This tool works with four shapefiles: ¼-mile and ½-mile forage partitions as polygon feature layers developed from the other RCWFAT tools, or any other method that correctly maps these partitions; a point feature layer of the center of each cluster; and a feature layer of stand polygons with an attribute table that contains data fields compatible with the 2003 Recovery Plan criteria as listed in Table 1. The fields can be named anything the analyst prefers as long as the name is not more than 10 characters. Table 1. System to score stands in relation to criteria specified in 2003 recovery plan using 12 stand characteristic criteria. For each criterion meeting the values specified for GQFH a score of 5 is assigned. Smaller scores are assigned to values of the stand characteristic further from the criterion. If all twelve criteria are measured a score from will be assigned to the stand. Score Stand Characteristic #14 + Pine stems < Basal Area 14 + pines < Basal Area pines > Basal Area < 10 pines > # pines < 10 > Basal Area of pine > 10 < % herbaceous ground cover < Hardwood midstory T-D M-M M-S L-M L-S Tall T (>15 ) Dense - D M-D T-S L-D Medium M (7-15 ) Medium - M T-M Low L (<7 ) Sparse S (Hardwood pulpwood BA) > <10 Stand Age Fire return interval or 4 <3 Fire type (Season of burning ) NGS GS 141

4 The tool uses a simple inquiry box with drop-down list to allow the analyst to declare which shapefiles in the map document have the required data. Once the correct shapefiles have been chosen, a field identity and initial evaluation form (Figure 1) allows choice of which fields in the stand feature layer have the necessary attributes, in the correct format, to evaluate the criteria in Table 1. Each criterion appears as a question as to which field contains data for that criterion, and a list of the fields in the stand feature layer are displayed in Label 2. Once a field is chosen and verified, the next criterion is displayed. The program will allow the analyst to skip ground cover and the two burning criteria and compensate for their lack with a reduced maximum possible score. If any of the others are absent, execution will stop and a message will be displayed that the fields are required for the program to operate. In addition, a field is also requested for stand species type to separate potential forage from hardwood and other non-forage types (roads, fields, lakes etc.). Once all the appropriate fields are entered, the three buttons along the bottom of the form (Figure 1) become active. The Initial Forage Evaluation button performs three actions, the stand layer and ½-mile forage partition are unioned to produce a new stand layer with the cluster number assigned to all stands and stand fragments within a foraging partition. Then all stands are evaluated by the criteria in Table 1 to establish a stand score. This score is multiplied by acres and divided by the maximum stand score to establish the Equivalent Good Quality Foraging Habitat in that stand. This value is rounded to a whole number and recorded as the W/L management index as an integer. The program adds a number of fields to the attribute table and it is permanently added to the map document and workspace for future use. The Select a Set to Export button is then used to choose the subset of stands for evaluation. At present this button executes a select-by-attribute dialog allowing the user to select any subset, or the whole data set, of the stands in the evaluated data for management impact in a what if session (Figure 2). The selected set is extracted into a new feature layer leaving the initial evaluation unchanged. After the analyst selects a subset and executes the Export Selects for Impact Evaluation button (Figure 1), a new dialog box appears (Figure 2) with attributes of the first selected stand already listed in Labels and the file pointer on the first record so it can be updated with proposed changes resulting from the first what if session. The dialog box where the what if sessions take place (Figure 2) consists of four sections. The upper seven labels contain stand identification information and the present score that remain constant as well as an update score which changes after alterations are made. The center lists eight criteria which will change with management options and the present value of each criterion (Labels 22-29) as current status. Right of these current status boxes are editable boxes which initially match the current status values. Management options are entered by editing the values in these boxes to match the results of the proposed management. Since several factors are interrelated, the program maintains mathematical integrity between these stand parameters. That is, if stems or basal area of >14 or basal area trees are changed the basal area >10 automatically is set to the new sum. It also maintains the initial relationship between basal area and stems by automatically changing either basal area or stems when the other is edited. The user is advised via message box when the changes are about to occur. There is a minimum 142

5 threshold that defines a pine stand as red-cockaded woodpecker habitat. If the user chooses to override that minimum (e.g. clear the stand) the stand is considered non-habitat and all points, initial and added, for that stand are lost from the management index score. Figure1. Dialog box. The user identifies the fields containing the necessary data for each parameter, runs the initial evaluation, selects stands from the evaluated data set, and exports the selected stands to a new feature layer for what if impact evaluation. Within the framework of data consistency, the user can propose leaving any number of stems or basal area for any parameter; however none can be added because there is no provision for growth in this model. When the user has proposed changes, the UpDate Status command will show the impact of the proposed action in two places. The updated score (Label 57) will be displayed and the current index, proposed index, amount of change and direction are displayed in Labels and 47. At this point the proposed change is not permanent and the user may edit the proposed criteria values and press the UpDate Status button again. This process can be repeated until the user is satisfied with the result. Just to the right of the proposed status boxes is a large area labeled GQFH deficiencies. The reasons this stand does not meet the criteria of Good Quality Foraging Habitat are listed in this box as an aid to the user. Once the user is finished with the stand, the Next Stand button is pressed which resets the form and pointer to the next stand in the file. It also makes the edits of the first stand permanent in the selected feature layer. As more stands are edited the cumulative results are listed in Labels 37-41and

6 Figure 2. Example of data form used for management evaluator. Each stand in the chosen set is evaluated separately. Proposed actions are entered in the right column by altering the value. Only possible alterations are allowed. The UpDate Status button calculates the new stand score and index value. This button can be used repeatedly, until the analyst is satisfied. The Next Stand button records the change and updates the cumulative box at the bottom. The other four buttons are used to report results. 144

7 Reporting can also take place at different levels and in different forms. In addition to the information being reported on the screen, printed and file reports can be generated for the whole selected set, forage partition subtotals, or every stand individually in the selected data set. A screen dump for any individual stand is also possible. This is possible because a new feature layer of the selected set has been created and is where all the updates for proposed changes take place. This exported feature layer can also be used in ArcMap or other software for additional analysis or graphical display. This also preserves the integrity of the initial evaluation data set so the user can return and select another subset for impact analysis. DISCUSSION In this initial development stage of the tool, we used it to evaluate the utility of the equivalent good quality foraging habitat index as a forest management planning tool. We were looking for an index with enough sensitivity to indicate change at the smallest management units used in forestry. We wanted the index scores to be cumulative in order to evaluate impacts to multiple management units. We wanted an index that would tie quality of habitat to quantity of habitat in each management unit. The resulting index was designed to work as an extension to any of the base scoring systems being proposed for foraging habitat. However the tool is developed only for the simplest one at this time. We used the tool in combination with the proposed index as an initial model to identify the needed inputs and outputs for such a system. The criteria in Table 1 are a mixture of stand characteristics that are subject to direct manipulation (stems and basal areas), subject to indirect manipulation (herbaceous ground cover), and are fixed (age). It also includes two fire management manipulations that are used to change groundcover, but are not included in the habitat characteristics listed in the Recovery Plan. At present the tool is designed to evaluate stand characteristics that can be directly manipulated. We use vegetative characteristics that are recorded as a part of forest inventory and are in historical data sets. Fire history is not usually included in forest stand records nor is groundcover. As a model, the program and index revealed some unforeseen constraints. The relationship between basal area, stems, and cumulative basal area must be maintained while proposing changes to any components. At present the relation between stems and basal area are set to remain exactly as those in the input data. This does not allow representation of cuts that change average stem diameter. The need to maintain a minimum threshold to define a stand as RCW foraging habitat was discovered because under some initial conditions it was possible to clear cut the stand and achieve an increase in stand score. Notice in Table 1, a stand will get ten points for less than ten square feet of pines under ten inches and hardwood pulp, including zero. In fact, a clear-cut stand that was burned every two years in the growing season would get at least 20 of the possible 60 points. Working with stand data that is usually available, only management that involves removal of vegetation, to change the size and species composition of stands, can be evaluated with this system. 145

8 CONCLUSIONS The index developed in this effort does overcome one large problem found when trying to apply the 2003 Recovery Plan to forest management. Very little of the forest now supporting woodpeckers meets the habitat guidelines of the Good Quality Foraging Habitat. On most of this land, a single silvicultural treatment will not result in solution of all the shortcomings. The model proposed here provides both a score of suboptimal stands, and an index to combine these scores into a single metric of habitat quality. Computation of this metric will determine if proposed silvicultural treatments (presently those that remove vegetation) result in habitat nearer to, or farther from, the Recovery Plan definition of Good Quality Foraging Habitat. In addition to the practical utility of this metric, it also provides a metric to test the actual performance of Good Quality Foraging Habitat to increase RCW populations. These criteria were developed to describe an ecosystem that provides habitat for observed successful RCW populations, as were those in the 1985 plan. The limitations of the criteria included in the 1985 plan were not recognized until the late 1990 s (Beyer et al., 1996; Wigley et al., 1999). The index developed here can be computed for single clusters to be compared to population metrics such as number of fledges. It can also be computed for a number of clusters to compare to cluster change such as abandonment or budding. Finally it can be calculated for entire forests to compare to actual woodpecker population changes. High index values, reflecting habitat close to Good Quality Foraging Habitat, should be closely correlated to measures of population gain. ACKNOWLEDGEMENTS We would like to thank Cynthia Ragland, and Jim Shores for cluster center and stand data used in index calculations. We would also like to thank Ralph Costa for sharing preliminary tables of habitat scoring. REFERENCES Beyer, D.E., R. Costa, R.G. Hooper, and C.A. Hess Habitat quality and reproduction of red-cockaded woodpecker groups in Florida. Journal of Wildlife Management. 60: Henry, V.G Guidelines for preparation of biological assessments and evaluations for the red-cockaded woodpecker. Atlanta, GA: U.S. Fish and Wildlife Service, Southeast Region, USA. James, F.C., C.A. Hess, B.C. Kicklighter, and R.A. Thum Ecosystem management and the niche gestalt of the red-cockaded woodpecker in longleaf pine forests. Ecological Applications. 11: Lipscomb, D.J., and T.M. Williams. In press. Evaluating some proposed matrices for scoring sub-optimal red-cockaded woodpecker foraging habitat in relation to the 2003 recovery plan. 13th Biennial Southern Silviculture Conference, USDA, For. Ser., South. Res. Stat. Lipscomb, D.J., and T.M. Williams Development of a geographic information system 146

9 based model of red cockaded woodpecker habitat. In Red Cockaded Woodpecker: Road to recovery, , Costa R.A. & S. J. Daniels, eds. Blaine WA: Hancock House. Wigley, T.B., S.W. Sweeney, and J.R. Sweeney Habitat attributes and reproduction of redcockaded woodpeckers in intensively managed forests. Wildlife Society Bulletin. 27: Wood, G.W., L.J. Niles, R.M. Hendrick, J.R. Davis, and T.L. Grimes Compatibility of even-aged timber management and red-cockaded woodpecker conservation. Wildlife Society Bulletin. 13: U.S. Fish and Wildlife Service Red-cockaded woodpecker recovery plan. Atlanta, GA: U.S. Fish and Wildlife Service, Southeast Region. U.S. Fish and Wildlife Service Recovery plan for the red-cockaded woodpecker (Picoides borealis), Second Revision. Atlanta, GA: U.S. Fish and Wildlife Service. 147