Mapping Black-tailed Prairie Dog Mapping Black-tailed Prairie Dog Colonies Across Montana Using the Colonies Across Montana Using the National Agriculture Imagery National Agriculture Imagery Program (NAIP) 2005 Imagery Program (NAIP) 2005 Imagery Bryce Maxell, Interim Director / Senior Zoologist (406) 444-3655 [email protected]Scott Blum, Biologist / Information Specialist Karen Walker, Biological Data System Coordinator March 13 March 13 th th 2010, Great Falls, Montana 2010, Great Falls, Montana http://mtnhp.org
49
Embed
Mapping Black-tailed Prairie Dog Colonies Across Montana Using the National Agriculture Imagery Program (NAIP) 2005 Imagery Bryce Maxell, Interim Director.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Mapping Black-tailed Prairie Dog Colonies Mapping Black-tailed Prairie Dog Colonies Across Montana Using the National Agriculture Across Montana Using the National Agriculture
Imagery Program (NAIP) 2005 ImageryImagery Program (NAIP) 2005 Imagery
Bryce Maxell, Interim Director / Senior Zoologist(406) 444-3655 [email protected]
Scott Blum, Biologist / Information Specialist
Karen Walker, Biological Data System Coordinator
March 13March 13thth 2010, Great Falls, Montana 2010, Great Falls, Montana
http://mtnhp.org
Montana’s Conservation Plan forBlack-tailed and White-tailed Prairie Dogs
• Objective #2 - develop statewide and regional prairie dog distribution and abundance standards
- Inventory and monitor distribution and abundance
- 7 km complexes of 5,000, 1,000-5,000, and <1,000 acres- 1.5 km complexes of 5,000, 1,000-5,000, and <1,000 acres (Biggins et. al. 2006)
• Objective #3C – identify isolated praire dog colonies in need of special consideration, assess their needs, and implement special management tasks, as appropriate
Datum Shift?
Weaknesses in Current Spatial Representation
Missed Town
Weaknesses in Current Spatial Representation
• Shifted PolygonsMissed ColoniesGround Mapping Limitations (can’t see the town for the burrows)
Weaknesses in Current Spatial Representation
National Agriculture Imagery Program
• Administered by USDA Farm Service Agency• 4 bands of information (red, green, blue and near
infrared) gathered during growing season in continental U.S. to yield both natural color and color infrared imagery
• Acquired at 1-meter ground sample distance or resolution (i.e. each pixel = 1 square meter)
• Horizontal accuracy = 6 meters• Montana first flown in 2005 (available in 2006)• MT flown again in 2009 (available March 4, 2010) • Add 2009 NAIP to ArcMap under ArcGIS Server
• First step toward evaluating use of NAIP imagery for digitizing prairie dog colonies
• Statewide spatial representations for areas with recent evidence of prairie dog activity for environmental reviews
• Summaries of colonies and complexes by various administrative units
• Identification of complexes for management and potential Black-footed Ferret reintroduction using both 7 km and 1.5 km rules
Methods 1• Use existing point and polygon data to create a
Maxent predictive distribution model• Stratify a network of 717 20 km x 20 km grid tiles
into high, medium, and low density areas• Test mapping effort on a random selection of 92
20 km x 20 km grid tiles• For each tile, blindly evaluate 1 hectare grid cells for
evidence of recent prairie dog activity (code = 1) using 2005 NAIP color and IR, MaxEnt model, slope model, and black and white DOQQ imagery at scales between 1:5,000 and 1:30,000 (1:10:000 seemed to work best)
• Turn on existing point, polygon, and flight data and evaluate against grid cells turned on
• Revaluate areas without recent evidence recorded and if evidence is detected (code = 2)
Maxent Model for Black-tailed Prairie Dogs
Black Points = current point observations and colony centroidsModel is a logistic surface varying from 0 to 1. Warmer colors predict more suitable habitat
Maxent Model Used to Develop Density Strata(Red = High, Green = Low, Blue = No Previous Documentation)
* Grid Tiles are 20 km x 20 km
Density Strata Relative to Existing Point Data (Red = High, Green = Low, Blue = No Previous Documentation)
* Grid Tiles are 20 km x 20 km
Pilot Mapping on Random Grid Cells (blue outline)
* Grid Tiles are 20 km x 20 km
Code = 1 for 1 Hectare Grid Cells on each Tile with Evidence of Recent Prairie Dog Activity
Compare Grid Cell Coding Versus Points, Polygons, and Flight Lines at 20x20 km scale
Compare Grid Cell Coding Versus Points, Polygons, and Flight Lines at 20x20 km scale
Code = 2 for Grid Cells That Second Review shows Evidence of Recent Activity
Missed on Initial Pass
Slight Discrepancy with 2008 Flight Line
Flight Line, but No Burrows Evident
Odd Mottling Not Coded as Recent Activity
Beware of Hay Bales
Initially Looks Possible, but on Side of Hill
Burrows?
Burrows?
Difficult Landscape 1
Difficult Landscape 2
Difficult Landscape 3
Problems with IdentifyingWhite-tailed Prairie Dog Colonies
What are these white features?
Methods 2• Stitch tiles together• Dissolve adjacent grid cells into colonies• Apply 750, 1,500 and 3,500 meter buffers to
identify 1.5, 3, and 7 km complexes• Calculate total area of complex• Code complexes by BLM Field Office, FWP
Region, Tribal Reservation, Forest, and Stewardship
• Evaluate complexes relative to Goals in Montana’s Conservation Plan
• NEED TO EVALUATE ACCURACY WITH GROUND TRUTHING!
Example for 1.5 and 7 km Complexes
Potential Colony and Complex Summary• 8,852 potential colonies (2,598 of these corroborated with previous info)
• 289 of the 7 km complexes- 6 ≥ 5,000 acres- 16 between 1,000-5,000 acres- 267 ≤ 1,000 acres- smallest = 2.5 acres, largest = 397,086 acres
• 1,032 of the 3 km complexes- 16 ≥ 5,000 acres- 42 between 1,000-5,000 acres- 974 ≤ 1,000 acres- smallest = 2.5 acres, largest = 109,883 acres
• 2,474 of the 1.5 km complexes- 10 ≥ 5,000 acres- 73 between 1,000-5,000 acres- 2391 ≤ 1,000 acres- smallest = 2.5 acres, largest = 38,766 acres
Note: Acres have not been verified with ground truthing andare certainly an overestimate ofon-the-ground acres due to the 1 Ha mapping unit used in thiseffort and likely misidentificationof other features as colonies.
NAIP (Black) Old Points (Purple) 2008 Flight (Red)
Colony Stewardship(Private = Gray, Federal = Yellow, State = Blue, Tribal = Brown)
6 x >5,000 Acre 7 km Complexes
7 km 5,000 Acre Complex Stewardship(Private = Gray, Federal = Yellow, State = Blue, Tribal = Brown)
16 x >5,000 Acre 3 km Complexes
3 km 5,000 Acre Complex Stewardship(Private = Gray, Federal = Yellow, State = Blue, Tribal = Brown)
10 x >5,000 Acre 1.5 km Complexes
1.5 km 5,000 Acre Complex Stewardship(Private = Gray, Federal = Yellow, State = Blue, Tribal = Brown)
Colony Size Class Distribution
7 km Complex Size Class Distribution
3 km Complex Size Class Distribution
1.5 km Complex Size Class Distribution
Summary byFWP Region
Region No. Colonies Total Acres
2 1 77
3 143 12,414
4 525 41,953
5 1,581 110,208
6 1,459 143,458
7 5,184 337,979
Total 8,893* 646,089**Totals are greater than total number of colonies and total acreage because some colonies are split between FWP Regions.
Summary byBLM Field Office
Field Office No. Colonies Total Acres
Billings 1,558 104,801
Butte 109 10,186
Dillon 31 2,164
Lewistown 544 43,933
Malta 1,364 142,534
Miles City 5,268 342,471
Total 8,893* 646,089**Totals are greater than total number of colonies and total acreage because some colonies are split between BLM Office.
*Some colonies are split between counties so sum of colonies and acreages here will exceed total no. & acreage of colonies.
Isolated Colony Near Grey Cliff SE of Big TimberConservation Issues? Opportunities for Research?
Suggested Uses and Future Work• Do not use acreage values from this assessment• Evidence of recent activity ≠ presence• Use with previous points and polygons when possible• Use polygonal layer of colonies in environmental reviews
and broad scale planning, but note the need to evaluate individual polygons on the ground
• Not useful for White-tailed Prairie Dog colonies• Ground truth random selection of previously unconfirmed
colonies to evaluate overall mapping accuracy (2005 and 2009)
• Ground truth unconfirmed colonies outside current known range
• Ground truth unconfirmed colonies on upper Missouri upstream of Craig and those that are very isolated
• Ground truth and update mapping for portions of large complexes (use 2009 NAIP and GPS info)