INTRODUCTION

This report summarizes work completed to date on an effort to correlate habitat and field size changes along mourning dove (Zenaida macrura) call count routes in the Western Management Unit (WMU) with changes in dove density recorded along these routes. We are using a GIS-based analysis of black/white aerial photography, obtained primarily from the U.S. Department of Agriculture (USDA) and the U.S. Geological Survey (USGS), to quantify for the 1960's, 1970's, and 1980's, the extant (hectares) of habitats included within one mile either side and from the ends of call-count routes (2 x 22 mile area). We selected a sample of 15 routes from several states of the WMU. We also conduct the analysis within a 1/2 mile radius of the 20 listening points along each route. We have obtained basic dove density information from the Office of Migratory Bird Management (MBMO), including route regression trends and number of doves heard. In the final analysis, we will compare documented extent of habitat changes along the routes with declines or increases in dove density on he routes o determine if here are patterns of specific kinds of habitat losses or increases, or field sizes, that correlate with dove density. This information can be used to explain the long-term decline of mourning doves in the WMU and should be useful to land managers to improve habitat conditions to increase dove production.

PERSONNEL

In June, Derik Ogden, working as an employee of the California Department of Fish and Game (CDFG), assisted in compiling data to be used for selection of routes for analysis. He compiled information on route trends (slope), average number of doves heard, years of data available for each route, number of missed years, number of observers, number of years with 0 doves heard, high and low range of doves heard, and stratum for each route. Derik also identified sources in individual states that potentially could supply aerial photography in the event that USDA and USGS could not.

In August, Blankenship and Miller interviewed potential GIS technicians and selected Chris Gregory who began work in October. Chris is on CDFG payroll, but USGS-BRD will hire him this summer for 3 months, to meet CDFG hiring rules, after which, he will return to CDFG.

Casey Stemler, originally assigned 40% to this project, left the Dixon Field Station to take a permanent position in the Biological Resources Division central office in Reston, VA. The Project Leader, Miller, has taken over his duties on this project.

EQUIPMENT/SUPPLIES

In July, we purchased computer hardware, a drive (X5204A 4.2GB 5400 RPM F/W SCSI-2) and additional memory (X116P 16MB SS20, SS10) from Sun Express, Inc., to store and manipulate the GIS data (see budget). We have purchased various photo indices, USGS quad maps, mylar overlays, pens, etc. over the last several months.

ROUTE SELECTION

We obtained route-regression trends and density data for each mourning dove call-count route in the WMU from MBMO. Using trend estimates, we sorted all routes in order of decreasing slope, and categorized routes into those in which doves have increased, decreased, and not changed between 1966 and 1987. In general, we wanted routes for analysis that: 1) had the steepest increasing or decreasing, or least changing, slopes with the smallest standard errors (error results from data variation and number of observers); 2) had the largest z-statistic (slope divided by standard error) ( 100; 3) had the largest number of doves on the route (( 8); and 4) had the fewest number of observers possible, but (6 (>3 per year) in any case. Selected routes had to have few if any zeros (no doves heard) unless zeros terminated long-term declines or preceded long-term increases. We rejected routes in which population density from year to year varied wildly, routes that did not have at least 17 years of data (out of 22) covering all 3 time periods. Within the WMU, we initially selected for continuing analysis the 5 routes in which doves decreased the most (greatest negative slope > 5%), the 5 routes in which doves increased the most (greatest positive slope > 5%), and the 5 routes in which doves changed the least (slope nearest 0, ( 3%) since 1966 when counts began.

Unfortunately, not all initially selected routes could be used because few of them met all the criteria. Finding replacements has proved difficult and time consuming. We relaxed certain criteria to include routes that seemed to show dramatic trends, but had fewer than the 8 minimum average number of doves. This occurred when the route started out with many doves being heard, then within a few years declined to few and then zero doves heard - the marginal mean is the mean number of doves during the "mid-years" (MBMO definition), so the mean gives a somewhat erroneous picture of the number of doves expected along the route. We accepted these kinds of routes only if the number of observers was (3. In general, we found that most routes had too many different observers over the time period to support trend analysis. For example, the California routes initially selected based on steep negative slopes were ultimately rejected because they had from 9-12 observers during the 22 years, which yielded large standard errors and low z-statistics. Ultimately, we had to settle for fewer than the objective number of 5 "Increasing" routes, because only 9 were available in the entire WMU, and only 2 met the minimum criteria for inclusion. In contrast, there was no shortage of "Decreasing" and "No change" routes. Specific routes may change again if we find that photography is not available for one or more of the time periods we are studying. Routes selected are in Table 1.

PHOTO ACQUISITION

We spent July and August obtaining route maps for selected and alternate routes from the WMU state fish and wildlife agencies. We received the last of these maps from Idaho and Nevada (late requests) in October. We used these to determine Lat-Long locations for ordering photos. We mailed requests for photography searches to the U.S. Geological Survey (USGS) Eros Data Center in Sioux Falls, SD, and to the U.S. Department of Agriculture (USDA) in Salt Lake City, UT on October 25, 1996. On November 4, we received a list of available USGS aerial photography from the Eros Data Center. This was a quick turn around, but it turned out to be incomplete and we had to request an additional search. This was also completed quickly and we received an update in late November. The USDA took longer to obtain photo lists, but they did more of the work for us by trying to limit photo lists to those that fell exactly within the area within one mile on either side of each route. The USGS search was less focused. On January 23, we sent in our first order for actual photographs after spending a good deal of time studying photo indices to make sure we would be getting the exact photos we needed. To date, we have not received any photography, and we don't expect any for at least one month.

PHOTO INTERPRETATION/DIGITIZING

Chris Gregory has interpreted and digitized habitat types for California route 120, after obtaining the photos from CDFG used in their analysis (see study proposal). Photography only covered 1/2 of the route during the 1960's, however, and we have ordered the remaining pictures. Chris produced color maps of the route for the 1960's, 1970's, and 1980's, and these are attached (Figure 1-3). Habitat types used for photo interpretation and to prepare the color maps are listed in Table 2. Based on Chris' experience with route 120, he estimates that from 2-10 days per route per time period will be required to complete all steps necessary to fully interpret and digitize habitats. The more complex the habitat mosaic, the more time that will be required. As Chris gains experience, the time may be reduced substantially, and we anticipate no problems in finishing the project on time provided that photography is forthcoming expeditiously. The entire process includes choosing correct photos, preparing photos for interpretation, photo interpretation, computer set-up, digitizing photos, cleaning of digitized images, converting to real-world coordinates, merging of images, adding attribute data, habitat analysis, final map creation, log creation, and other miscellaneous tasks.

RESULTS

We have preliminary habitat results for California route 120 for the 1970's and 1980's period (Table 3). The 1960's period was not completed because of the absence of photography. In addition to analysis of habitats within one mile of the call-count route, we decided also to do an analysis within 1/2 mile of each listening point along the route. This reduces the area and, in the future, will allow an analysis of habitat vs dove density, point by point along the route. This will allow us to determine if certain parts of any given route most influences the trend for that route. From the 1970's to the 1980's, our results suggest that there has been a sizable loss of water areas, riparian habitats, agriculture, and closed canopy woodland, but orchards, grasslands, and urban vegetation have increased substantially during the same period. These same trends are evident whether the 1-mile or 1/2-mile areas are considered. The overall effect that these changes could have on mourning dove nesting populations is unclear, but it is generally believed that losses of riparian habitat would eliminate nesting sites and the elimination of agricultural lands would reduce food resources. Changes in water areas might be misleading depending upon the time of year the respective photos were taken. The importance of closed canopy woodland to dove nesting is not known, but it is probably not as important as the open canopy woodland which, on this route, is very extensive and changed very little between the two time periods.

FUTURE WORK

We have ordered most of the photography for selected routes and photo searches for alternate routes. We will refine our analysis of California route 120 as we wait for delivery of these products. One of the problems we have encountered is that different scales of photography are available for each time period. Thus, small patches of habitats, such as water bodies, roads, urban vegetation, etc., and the distinction between open canopy woodland and grassland are perceived differently at different scales. We will attempt to standardize photo scale by adjusting photo size (e.g., ordering larger photos for photography available only in a small scale). Ideally, the 1:20,000 scale is most desirable. The 10" x 10" size of photo works well with this scale, but 1:40,000 and 1:60,000 scales will require larger sized photos to better standardize photo interpretation. Another potential problem is that some black/white photos will be prints from CIR film. This might yield lighter prints than from black/white film and affect comparative photo interpretation. We anticipate delivery of ordered photos by March.

BUDGET

Total expenditures to date include $8960 for the salary of Casey Stemler (Federal share) and $6134 for Chris Gregory (CDFG share). In addition, we spent $2497 on computer hardware, $796 on USGS quad maps, $6 for pens and mylar, and $38 on photo indices for aerial photography. We have ordered the first batch of aerial photography at projected costs of $827 from USDA and $266 from USGS.

_______________________________________________________________________________________________

Table 1. Basic route data for GIS analysis of habitat trends vs dove density in the Western Management Unit.
(Routes may change depending upon availability of aerial photography)
_______________________________________________________________________________________________

Increasing Number of Doves

State	Route	Slope	S.E.	Z-stat	Marg. Mean	Years of Data	Number of Observers	Strong Trend
AZ	1550	1.509	.009	167.7	9.8	21	6	n.a.
AZ	220	1.499	.010	149.9	27.7	21	6	n.a.

Decreasing Number of Doves

State	Route	Slope	S.E.	Z-stat	Marg. Mean	Years of Data	Number of Observers	Strong Trend
CA	130	0.792	.002	396.0	5.9	19	2	yes
ID	2520	0.835	.003	278.3	2.6	19	4	yes
CA	100	0.834	.003	278.0	15.1	19	2	n.a.
WA	150	0.885	.004	221.3	12.3	19	4	n.a.
NV	720	0.900	.002	450.0	4.3	22	4	yes
CA	550	0.925	.002	462.5	5.7	21	4	yes
WA	151	0.930	.009	103.3	23.4	21	4	n.a.
CA	120	0.926	.002	463.0	47.9	21	2	n.a.

No Change in Number of Doves

State	Route	Slope	S.E.	Z-stat	Marg. Mean	Years of Data	Number of Observers	Strong Trend
CA	190	1.004	.004	251.0	15.2	21	4	n.a.
AZ	60	1.000	.008	125.0	25.9	22	5	n.a.
UT	2830	1.000	.002	500.0	17.3	22	3	n.a.
CA	450	0.998	.001	998.0	53.6	22	1	n.a.
CA	520	0.996	.008	124.5	9.0	19	4	n.a.

_______________________________________________________________________________________________

Table 2. Habitat types used in GIS analysis of mourning dove call-count route habitat change 1960's - 1980's.
(Modified from Gates, R.J., J.L. Roseberry, and Alan Woolf. 1995. Mourning Dove Breeding Habitat Status Adjacent to Federal Survey Routes. Cooperative Wildlife Research Laboratory, SIUC).
_______________________________________________________________________________________________

WATER ONLY - Any aquatic area large enough to be digitized and without including another habitat type while doing so. Examples include lakes, reservoirs, rivers, streams, ponds, stock ponds and marshes.

RIPARIAN 1 - Any tree line associated with water, roads or fields, or any hedgerow (straight line of trees or brush) large enough to be digitized.

RIPARIAN 2 - Mixed areas of the above two categories which are too small to be digitized separately (i.e., water under riparian canopy).

AGRICULTURE - Open areas which are being plowed or show signs of being plowed in the past.

ORCHARD - Tree canopy in agricultural rows.

GRASSLAND - Areas where tree canopies provide (relatively) less than 5% ground coverage. These areas must show no signs of plowing. Grazing land is included in this category.

OPEN CANOPY - Wooded areas where tree canopies provide (relatively) 5% - 80% ground coverage. Orchards and riparian areas are not included in this category.

CLOSED CANOPY - Wooded areas where tree canopies provide (relatively) greater than 80% ground coverage.

URBAN VEGETATION - Natural areas with human change AND immediately adjacent to urban areas. Examples include yards, sub-division lots under construction, golf courses, etc.

URBAN - Areas of man-made human existence. Examples include houses and other related structures, barns, silos, sheds, airports, etc. Roads are digitized as dimension-less straight lines. _______________________________________________________________________________________________

Table 3. Hectares of coverage by habitat type within 1 mile of mourning dove call-count route 120, California, and within 1/2 mile of listening points along route 120, and percentage change in habitat coverage between 1970's and 1980's.
_______________________________________________________________________________________________

----------------------------------------------Hectares - 1 Mile Coverage------------Hectares - 1/2 Mile Coverage------

Habitat Types	1970's	1980's	% Change	1970's	1980's	% Change
Water Only	84	46	-45	25	14	-36
Riparian 1	33	10	-70	27	10	-63
Riparian 2	434	393	-9	294	250	-15
Agriculture	1177	544	-54	624	245	-61
Orchard	26	66	+154	12	38	+217
Grassland	4017	5498	+37	1835	2501	+36
Open Canopy Woodland	4865	4492	-8	1807	1734	-4
Closed Canopy Woodland	713	193	-73	284	90	-68
Urban Vegetation	1	10	+900	1	8	+700
Urban	13	12	-	9	8	-11

_______________________________________________________________________________________________

Figures 1a-c. Maps of habitat types as determined using GIS analysis and plotting of interpreted 1960's, 1970's, and 1980's black/white photography of California mourning dove call-count route 120. (Note: 1960's not completed because of unavailability of aerial photography at the time this report was prepared. Photography to complete the analysis has been ordered). _______________________________________________________________________________________________