An Analysis of the Upland Goals Project:

1
An Analysis of the Upland Goals
Project Representativeness and Ground-Truthing
Ashley Apel, Bridget Dobrowski, Dave Panitz,
Erica Eisch
2
Our Two Research Questions

• Are protected lands representative of existing
  proportions of vegetation classes?

• Are protected lands representative of existing
  proportions of vegetation classes?

2. What percent of selected lands are not great
for conservation?
3
Whats conserved now and in the future
Currently Conserved
Future Conserved
4
Whats conserved now and in the future
Currently Conserved
Future Conserved
5
Land Cover Type Distribution
6
Does representation change with conservation?
7
Rank 4 No Improvement
Rank 3 gt 50
Rank 2 gt 75
Rank 1 gt 90
8
Our Two Research Questions

• Are protected lands representative of the natural
  proportions of vegetation classes?

2. What percent of selected lands are not great
for conservation?
2. What percent of selected lands are not great
for conservation?
9

• Are GIS vegetation layers accurate?

Valley Oak Forest
Cool Grasslands
Water
Desert Scrub
Coast Live Oak Forest
Redwoods
Coastal Scrub
Serpentine Hardwoods
10
Random Selection of Hexagons
UG Selected Hexagons
Our 199-hexagon sample
11
Random Hexagons Exported to Google Earth
12

• We split land cover into 14 types, grouped into 3
  classes of greatness

13

• ranked the coverage types within each hexagon

14
1st Forest2nd Rock/Dirt
15
1st Road2nd Ranch3rd Scrub
16

• and formulated our own selection rules

• IN all of the following
• Ranks 1-3 are YES or blank
• Doesnt abut a city

• OUT any of the following
• Hexagon abuts a city
• Rank 1 is "Road" or any of the "NO" class
• Both Rank 2 and Rank 3 are in the "NO" class

• MAYBE everything else
• Not abutting a city
• Not dominated by road, suburbs, agriculture, or
  industry (quarry, mine, mill, factory, dump)
• Containing at least some of the above or some
  structure

17

• But nobody likes MAYBEs

so we hand-sorted MAYBEs into the OUT and IN
categories, considering

• connectivity to adjoining natural landscapes
• fragmentation (due to buildings or roads)
• road surface and density

Some examples
18
1st Scrub/Chaparral 2nd Road
19
1st Grassland 2nd Windfarm 3rd Road
20
Initial Results
Sensitivity Analysis
Relax Abutting City Constraint
Include Ranches and Windfarms
Exclude Ranches and Windfarms
21
Take-homes

• GIS layers may not match up well enough with
  reality ground-truthing essential
• More representative overall, but avoid unintended
  changes to existing habitat proportions
• Alter MARXAN rules to better incorporate road
  density
• Peri-urban buffer to avoid leapfrog development
  and high conservation costs
• Considerable suburban sprawl outside mask

22
Questions?

• aapel_at_bren
• bdobrowski_at_bren
• dpanitz_at_bren
• eeisch_at_bren