Maggie Morris, Kevin Johnston, and Kate Grandison

Delineation of Mexican Spotted Owl Canyon Habitat

ABSTRACT

Designation of the Mexican spotted owl as a threatened subspecies requires monitoring population and managing habitat for longterm viability. Inventories have located the owls almost exclusively in steep- walled canyons that fall within an identified range of elevation, canyon width, and canyon depth. The GRID module of ArcInfo was used to delineate canyons and record width and depth through these canyons in order to provide researchers with the capability of interactively displaying areas within specified ranges and performing further statistical analysis.

BACKGROUND

In response to perceived threats to critical habitat and lack of adequate protective regulations, the Mexican spotted owl (Strix occidentalis lucida) was the focus of a petition to the U.S. Fish and Wildlife (USFWS) for listing as a threatened species in 1989. In November 1991, the USFWS published their proposal to list the southwestern subspecies as threatened. Formal listing as a threatened subspecies occurred on March 15, 1993 (USDI 1993). This designation requires monitoring population and managing habitat for longterm viability of the species throughout its range.

Mexican spotted owls (MSO) are widely distributed throughout their range but may occur as subpopulations restricted to habitat "islands"(Ganey 1988, Ganey and Balda 1989, Willey 1993, Rinkevich 1991). The southern Utah subpopulation represents the northwestern edge of the MSO distribution. Subpopulations in southern Utah appear dispersed by geographic barriers (e.g. Lake Powell) and collectively may form a metapopulation (Shaffer 1985).

The earliest recorded spotted owls in Utah were located in Zion National Park (Kertell 1977, Rinkevich 1991). The most northerly occurrence was in the Book cliffs of northeast Utah in 1958. The largest population of Mexican spotted owls in Utah occurs in Zion National Park where surveys have confirmed 17 nest sites (Rinkevich 1991, Willey pers. comm.). Current Mexican spotted owl records in Utah indicate that there are approximately 50 locations across southern Utah on BLM, Forest Service, and National Park Service canyon lands (Rinkevich 1991, Willey, pers. comm.).

The USFS has been conducting formal spotted owl inventories in southern Utah on the Fishlake, Dixie, and Manti-LaSal National Forests since 1990. Very few confirmed breeding pairs of spotted owls have been located on any of the National Forests in these studies (SWCA 1990, 1991, 1992, 1993, HDR 1990, 1991, 1992, 1993, 1994). Currently, only four MSO pairs have been identified using the southern Utah National Forests; however, several MSO pairs have been located on adjacent BLM and NPS lands (Rinkevich 1991, HDR 1990, 1991, 1992, 1993). Historically and currently throughout southern Utah, spotted owls have been found nesting only on ledges or small caves in steep-walled canyon (Kertell 1977, Rinkevich 1991, HDR 1990 1991 1992 1993, SWCA 1990 1991 1992 1993, Willey 1992 1993 ), which may contain intermittent streams and stringers of mixed conifer and/or deciduous multi-layered vegetation.

The inventories in southern Utah encompassed a wide range of habitat types, but located MSO almost exclusively in steep sandstone canyons (Rinkevich 1991, Willey 1993). Areas occupied by MSO in the canyon country may fall into two distinct habitat types: the "mesic" sites (i.e. Zion National Park, Texas Canyon, Hammond Canyon, etc ) are characterized by relatively humid conditions with nearby standing water or perennial streams, well-developed riparian zones, and higher elevation slopes with pockets of Douglas-fir, while the "xeric" sites (i.e. Canyonlands National Park, Dirty Devil River, Capital Reef National Park, etc.) are characterized by arid, relatively open habitats with little vegetation coverage dominated by pinyon pine and Utah juniper and desert shrubs (Brown 1982, Willey 1993). The occupied canyon habitats are structurally different from previously studied MSO sites to the south in Arizona (Ganey 1988). More specific information on adult and juvenile movement patterns and habitat ecology in the unique canyonland region has been gathered through a telemetry study in southern Utah (Willey 1992, Willey 1993).

ALGORITHM DEVELOPMENT

The National Forests of Southern Utah, including the Dixie, the Fishlake, and the Manti-Lasal, worked together to identify all MSO potential suitable habitat in southern Utah as a part of the process of developing a management plan. Each of these three National Forests have established Geographic Information Systems capabilities. The GRID module of ArcInfo GIS was used extensively to complete this project.

Initial criteria used to quickly identify potential suitable habitat included areas where slope exceeds forty percent and elevation is below eight thousand feet. Selected vegetation types were also used in the study, which was a simple binary coincidence model. However, it soon became evident that the criteria needed to be refined. The "slope exceeds forty percent" criteria did not serve to isolate "steep-walled canyons". The GRID algorithm to delineate canyons and record width and depth of areas within the canyons was developed in response to this problem. It not only addresses the immediate requirement of more precise potential habitat delineation, but will provide the wildlife biologists with the capability of interactively displaying areas within a specified range of canyon width and depth and performing further statistical analysis as owl locations are identified.

The required output from this project is two grids. The first grid covers the analysis area with each cell containing a value that identifies the canyon width at that cell location. Let us first define "canyon width" as "the length of a line drawn perpendicular to flow direction along a drainage from ridge to ridge". In general terms, the algorithm requires that we analyze each cell along every drainage, determine its flow direction, and count the number of cells along a perpendicular line to the ridge on either side of the drainage. This total number of cells is multiplied by cell width, and the product, canyon width, is stored in each cell.

IMPLEMENTATION

All grids required by the Canyon Delineation algorithm were derived from USGS 7.5' Digital Elevation Model data. The FILL command was used to eliminate sinks in the DEMs. Some scan lines were noted in the test data. Attempts to filter them out without introducing distortions have been unsuccessful to date. While recognizing these limitations, the DEMs provide the best available information for this project.

Two grids must be derived from the DEM data in order to run the Canyon Delineation program. The first input grid, STRMDIR, contains cells that define drainages. Each cell value indicates flow direction at that cell location. All other cell values are NODATA. Specific commands used to develop this input grid are as follows:

Input grid ELEV is Elevation lattice with sinks FILLED FDIR = FlowDirection ( ELEV ) FLACC = FlowAccumulation (FDIR) STRMDIR = con ( FLACC > 100, FDIR )

The second input grid, RIDGE, delineates the ridgelines between drainages. This is a binary grid with cell value of 1 for all cells coincident with ridgeslines. All other cells contain NODATA. Specific commands used to develop this grid are listed below. However, it should be noted that development of the RIDGE grid is still being refined. What we are looking for is cells with minimal slope that are "near" cells with steep slope.

     SLP = SLOPE ( ELEV, PERCENTRISE )

     FLACCFILTER = FocalMax ( FLACC, Rectangle, 5, 5 )

     SLPFILTER = FocalMax ( SLP, Rectangle, 8, 8 )

     SLPCHANGE = SLPFILTER - SLP

     RIDGE = con (  SLPCHANGE > 30           AND

                    SLP < 20                           AND

                    FLACCFILTER < 1000  AND

                    FLACC < 20, 1 )

The output grid, CANYONWIDTH, contains canyon width values. All cells not located within a canyon contain NODATA. Any cell that may be contained in more than one "canyon slice" contains the minimum canyon width value. CANYONDEPTH contains depth of canyon at each corresponding cell location.

Implementation is performed in four phases. During the first phase, the STRMDIR input grid is analyzed and the distance from the STRMDIR cell to each ridge is determined. Since MSO suitable canyon habitat has been identified as canyons with a total width from 200 to 1835 meters, we restrict the search for ridges to 1800 meters on each side of the drainage cell. The second phase of implementation consists of recording the canyon width and depth as attributes in the STRMDIR grid. In phase three, the CANYONWIDTH and CANYONDEPTH output grids are created. Pseudocode is provided for each of these three phases. The final step is to combine the CANYONDEPTH and CANYONWIDTH grids with other information to identify potential suitable habitat.

PHASE ONE:
Analysis of STRMDIR data

(The following process is performed within a DOCELL loop) For each cell in the STRMDIR grid

     If flowdirection is North or South.

     Count number of cells West to RIDGE

     Record number in output grid WCYNWIDTH. 

     Record depth in WCYNDEPTH

     Count number of cells East to RIDGE

     Record number in output grid ECYNWIDTH. 

     Record depth in ECYNDEPTH

     Else If flowdirection is East or West

     Count number of cells North to RIDGE

     Record number in output grid NCYNWIDTH.

     Record depth in NCYNDEPTH.

     Count number of cells South to RIDGE

     Record number in output grid SCYNWIDTH

     Record depth in SCYNDEPTH

     Else If flowdirection is Northeast or Southwest

     Count number of cells Northwest to RIDGE

     Record number in output grid NWCYNWIDTH

     Record depth in NWCYNDEPTH

     Count number of cells Southeast to RIDGE

     Record number in output grid SECYNWIDTH

     Record depth in SECYNDEPTH

     Else 

     Count number of cells Northeast to RIDGE

     Record number in output grid NECYNWIDTH

     Record depth in NECYNDEPTH

     Count number of cells Southwest to Ridge

     Record number in output grid SWCYNWIDTH

     Record depth in SWCYNDEPTH

     Endif

PHASE 2:
Compute Canyon Width and Depth

Fields are added from the temporary grids generated in phase one. The purpose of adding fields to STRMDIR is to make it necessary to shift only ONE grid, STRMDIR, in phase three described later in this paper.

Add the following fields to STRMDIR attribute table: width: To store total canyon width West, SW, South, SE, East, NE, North, NW: To store cellcounts depth: To store canyon depth

   

If STRMDIR is North or South

     STRMDIR.East = ECYNWIDTH

     STRMDIR.West = WCYNWIDTH

     STRMDIR.width =(  ECYNWIDTH + WCYNWIDTH ) * 30

     STRMDIR.depth = Minimum of ECYNDEPTH and WCYNDEPTH

Else If STRMDIR is East or West

     STRMDIR.North = NCYNWIDTH

     STRMDIR.South = SCYNWIDTH

     STRMDIR.width =  ( NCYNWIDTH + SCYNWIDTH ) * 30

     STRMDIR.depth = Minimum of NCYNDEPTH and SCYNDEPTH

Else If STRMDIR is NE or SW

     STRMDIR.NW = NWCANYON

     STRMDIR.SE = SECANYON

     STRMDIR.width = ( NWCANYON + SECANYON ) * 42.42

     STRMDIR.depth = Min of NWCYNDEPTH and SECYNDEPTH

Else If STRMDIR is NW or SE

     STRMDIR.NE = NECANYON

     STRMDIR.SW = SWCANYON

     STRMDIR.width = ( NECYNWIDTH + SWCYNWIDTH ) * 42.42

     STRMDIR.depth = Min of NECYNDEPTH and SWCYNDEPTH

EndIf

PHASE THREE:
Create CANYONWIDTH and CANYONDEPTH output grids

CANYONWIDTH = STRMDIR.width

CANYONDEPTH = STRMDIR.depth

Loop 60 Iterations to travel 1800 meters or more each direction:

     WEST = Shift STRMDIR grid to the left (  * 30 ) 

     If STRMDIR.West greater than 

           If NODATA value for CANYONWIDTH 

          or current value >  STRMDIR.width

                Update CANYONWIDTH and CANYONDEPTH

     EAST = Shift STRMDIR grid to the right (  * 30 ) 

     If STRMDIR.East greater than 

          If NODATA value for CANYONWIDTH

            or current value > STRMDIR.Canyonwidth

          Update CANYONWIDTH and CANYONDEPTH 

     NORTH = Shift STRMDIR grid up (  * 30 )

     If STRMDIR.North greater than 

     If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width 

          Update CANYONWIDTH and CANYONDEPTH

                    

     SOUTH = Shift STRMDIR grid down (  * 30 )

     If STRMDIR.South greater than 

          If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width

          Update CANYONWIDTH and CANYONDEPTH

 

     NE = Shift STRMDIR grid up and to the right ( * 30)

     If STRMDIR.NE greater than 

          If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width

          Update CANYONWIDTH and CANYONDEPTH

    NW = Shift STRMDIR grid up and to the left (  * 30)

     If  STRMDIR.NW greater than 

          If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width

          Update CANYONWIDTH and CANYONDEPTH

     SE = Shift STRMDIR grid down and to the right (  * 30)

     If STRMDIR.SE greater than 

          If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width

          Update CANYONWIDTH and CANYONDEPTH

     SW = Shift STRMDIR grid down and to the left ( < loop index> * 30)

     If STRMDIR.SW greater than 

          If NODATA value for CANYONWIDTH

     or current value > STRMDIR.width

          Update CANYONWIDTH and CANYONDEPTH

End Loop

When these three phases have been completed, interim grids can be deleted, leaving the final two output grids, CANYONWIDTH and CANYONDEPTH.

PHASE FOUR:
Determine Suitable Potential Habitat

The output grids CANYONWIDTH and CANYONDEPTH can be used to refine criteria for canyon structure when analyzed in conjunction with actual locations. The final task is to produce a grid describing potential suitable habitat would require overlaying CANYONWIDTH and CANYONDEPTH with vegetation, elevation, and slope data, and using a simple IF statement to select cells meeting criteria in all input grids.

CONCLUSION

Implementation of the algorithm described in this paper is ongoing and individual modules have been tested on a small test dataset. It will continue to be developed and refined and it is hoped that in the near future it will be used on real-world data from the southern Utah National Forests to assist in efforts to monitor the Mexican spotted owl population and manage habitat for longterm viability of the species.

REFERENCES

Brown, D.E. 1982. Biotic communities of the American southwest- United States and Mexico. Desert Plants 1-4. 342pp.

Ganey, J.L. 1988. Distribution and habitat ecology of the Mexican spotted owls in Arizona. M.S. Thesis Northern Arizona Univ., Flagstaff. 229pp.

Ganey, J.L. 1992. Food Habitats of Mexican spotted owls in Arizona. Wilson Bull. 104:321-326.

Ganey, J.L., and R.P. Balda. 1989. Distribution and habitat use of Mexican spotted owls in Arizona. Condor 91:355-361.

Ganey, J.L., R.P. Balda, and R.M. King. 1993. Metabolic and evaporative water loss of Mexican spotted owls and great horned owls. Wilson Bull. 105:645-656.

HDR 1990. 1990 Mexican spotted owl inventory: Final Report.

HDR 1991. Mexican spotted owl surveys in southern Utah: Final Report.

HDR 1992. Mexican spotted owl inventory and reinventory: Toiyabe, Manti-LaSal and Fishlake National Forests. Final report.

HDR 1993. Mexican spotted owl inventory and reinventory: Toiyabe, Manti-LaSal, Dixie, and Fishlake National Forests. Final Report.

Kertell, K. 1977. The spotted owl at Zion National Park, Utah. Western Birds 8:147-150.

McDonald, C.B., J. Anderson, J. Lewis, R. Mesta, A. Ratzlaff, T. Tibbits, and S. Williams. 1991. Mexican spotted owl, Strix occidentalis lucida: status review. Endangered species report 20. U.S. Fish and Wildlife Service, Department of Interior. Albuquerque, N.M. 85pp.

Rinkevich, s.e. 1991. Distribution and habitat characteristics of Mexican spotted owls in Zion National Park, Utah. Masters thesis, Humboldt State Univ.

Shaffer, M.L. 1985. The metapopulation and species conservation: the special case of the northern spotted owl. pages 86-99 In: R.J. Gutierrez and A.B. Carey [eds.] Ecology and management of the spotted owl in the Pacific Northwest. USDA For. Serv. Gen. Tech. Rep. PNW-185, Portland, Oregon.

SWCA 1990. Mexican spotted owl inventory: Dixie National Forest. Final Report.

SWCA 1991. Mexican spotted owl inventory and reinventory on the Dixie National Forest. Final Report.

SWCA 1992. Mexican and California spotted owl inventory: Dixie, Fishlake, Manti-LaSal and Toiyabe National Forests. Final Report.

SWCA 1993. Mexican and California spotted owl reinventory: Dixie, Fishlake, and Toiyabe National Forests. Final Report.

Willey, D. 1992. Habitat ecology of Mexican spotted owls in southern Utah. Utah Division of Wildlife Resources Technical Report.
Contract no.91-2577. 34pp.

Willey, D. 1993. Home-range characteristics and juvenile dispersal ecology of Mexican spotted owls in southern Utah. Utah Division of Wildlife Resources Technical Report. Contract No. 91-2577. 44pp.

Maggie Morris
GIS Coordinator
Dixie National Forest
82 North 100 East
Cedar City, Utah 84721
Telephone: (801) 865-3737
Fax: (801) 865-3791

Kevin Johnston
GRID Specialist
Esri
380 New York Street
Redlands, CA. 92373
Telephone: (909) 793-2853
Fax: (909) 793-5953

Kate Grandison
Mexican Spotted Owl Coordinator
Dixie National Forest
82 North 100 East
Cedar City, Utah 84721
Telephone: (801) 865-3739
Fax: (801) 865-3791