Engelhardt, Blake

Where in the basin are they?<br />

Validating habitat suitability models for 22 rare plant species.<br />

Stuart Osbrack<br />

Steve Matson<br />

USFS<br />

<strong>Blake</strong> <strong>Engelhardt</strong> & Shana Gross<br />

USFS - Lake Tahoe Basin Management Unit<br />

May 23, 2012

---

Background<br />

National forests are tasked with:<br />

1) maintaining viable populations of TES species<br />

2) preventing degradation of suitable habitat<br />

What constitutes suitable habitat?<br />

How much of it occurs on the landscape and where is it?<br />

LTBMU<br />

25 sensitive botanical species<br />

(19 vascular plants, 4 bryophytes, 1 lichen, 1 fungi)<br />

2 candidate species: Rorippa subumbellata (Tahoe Yellowcress)<br />

Pinus albicaulis (Whitebark Pine)*<br />

*Warranted but precluded from listing under ESA, July 2011

---

Binary GIS Models<br />

• Species distribution predicted from environmental variables<br />

• Classify landscape: suitable vs. non-suitable<br />

• Model Builder: Boolean AND/OR analysis of various<br />

environmental data (GIS layers) to determine suitable<br />

habitat<br />

Products<br />

• Total area of suitable habitat<br />

• Spatial pattern/location of<br />

suitable habitat

---

-15 habitat models, cover 22 species total<br />

-known or likely to occur in Lake Tahoe basin<br />

Wet/Riparian<br />

Botrychium spp.<br />

Bruchia bolanderi*<br />

Epilobium howellii<br />

Meesia spp./Helodium blandowii<br />

Peltigera hydrothyria*<br />

USFS<br />

Upland/Forested<br />

Arabis rectissima v. simulans<br />

Arabis rigidissima v. demota<br />

Arabis tiehmii*<br />

J. Morefield<br />

Eriogonum umbellatum v. torreyanum*<br />

Hulsea brevifolia*<br />

Lewisia kelloggii*<br />

Alpine<br />

Draba asterophora v. asterophora<br />

Draba asterophora v. macrocarpa<br />

Draba-Union<br />

Lewisia longipetala<br />

USFS

---

Example: Lewisia longipetala<br />

General Criteria GIS Dataset Refined Criteria (AND)<br />

Northerly Aspect Aspect 0-90°, 270-360°<br />

High Elevation Elevation >8300 ft<br />

Subalpine & Alpine<br />

Communities<br />

Potential Natural<br />

Vegetation<br />

TSME, TSME-ABMA, Rocky,<br />

Scree-subalpine/alpine<br />

Granitic soils Geology Igneous intrusive, and<br />

extrusive, sedimentary

---

Botrychium spp.<br />

(Moonwort complex)<br />

Peltigera hydrotheria<br />

(Veined Water Lichen)<br />

Draba asterophora v. asterophora<br />

(Tahoe Draba)<br />

Lewisia kelloggii<br />

(Kellogg’s Lewisia)

---

ARRE<br />

ARRID<br />

ARTI<br />

BOTRY<br />

BRBO<br />

DRABA UNION<br />

DRASA<br />

DRASM<br />

EPHO<br />

ERUMT<br />

HUBR<br />

LEKE<br />

LELO<br />

MEESIA_HEBL<br />

PEHY<br />

Lake Tahoe Basin<br />

(hectares)<br />

Predicted Area of Suitable Habitat w/in Lake Tahoe Basin<br />

90000<br />

80000<br />

70000<br />

60000<br />

50000<br />

40000<br />

30000<br />

20000<br />

10000<br />

0<br />

(1 hectare = ~2.5 acres)

---

Field Validation<br />

• 500m x 500m grid<br />

• 100 randomly selected points<br />

• 15 excluded (private property)<br />

• 85 total points<br />

• Field visit: presence/absence<br />

of suitable habitat & species<br />

• 45 of 85 completed<br />

2009- 9 points<br />

2010- 4 points<br />

2011- 32 points<br />

2012- 40 points planned… any volunteers?!

---

45 Visited Points:<br />

Few points on North Shore<br />

All elevations<br />

All species

---

40 Remaining Points:<br />

Throughout the basin<br />

All elevations<br />

All species

---

Preliminary results:<br />

- 7 of 15 models had >90% accuracy<br />

- mostly over-prediction<br />

- all upland and alpine species<br />

Species<br />

% Sites<br />

Habitat<br />

Predicted<br />

% Model<br />

Agreement<br />

% Overpredict<br />

% Underpredict<br />

Habitat<br />

Arabis tiehmii 2.2 100 0 0 Upland<br />

Draba-Union 3.1 97 0 3 Alpine<br />

Draba asterophora v.<br />

macrocarpa<br />

Draba asterophora v.<br />

asterophora<br />

13.3 96 4 0 Alpine<br />

15.6 93 7 0 Alpine<br />

Lewisia longipetala 15.6 93 7 0 Alpine<br />

Arabis rigidissima v.<br />

demota<br />

Eriogonum umbellatum<br />

v. torreyanum<br />

57.8 91 9 0 Upland<br />

11.1 91 7 2 Upland

---

Preliminary results:<br />

- 4 of 15 models had 80-89% accuracy<br />

- all over-predict, some under-prediction<br />

- riparian and upland species<br />

Species<br />

% Sites<br />

Habitat<br />

Predicted<br />

% Model<br />

Agreement<br />

% Overpredict<br />

% Underpredict<br />

Habitat<br />

Botrychium spp. 86.7 89 11 0 Wet/Riparian<br />

Arabis rectissima v.<br />

simulans<br />

28.9 82 9 9 Upland<br />

Bruchia bolanderi 84.4 82 16 2 Wet/Riparian<br />

Hulsea brevifolia 62.2 80 20 0 Upland

---

Preliminary results:<br />

- 4 of 15 models had low (67-71%) accuracy<br />

- most over-predict AND under-predict<br />

- mostly riparian species<br />

- missing some mesic habitats<br />

Species<br />

Peltigera<br />

hydrothyria<br />

% Sites<br />

Habitat<br />

Predicted<br />

% Model<br />

Agreement<br />

% Overpredict<br />

% Underpredict<br />

Habitat<br />

57.8 71 20 9 Wet/Riparian<br />

Lewisia kelloggii 73.3 69 29 2 Upland<br />

Epilobium howellii 48.9 67 4 29 Wet/Riparian<br />

Meesia spp. &<br />

Helodium blandowii<br />

46.7 67 16 18 Wet/Riparian

---

What’s working…<br />

• Elevation, aspect, slope, high<br />

elevation forests<br />

(Draba, Lewisia)<br />

• Conifer/forest types- generalists<br />

(Arabis, Eriogonum)<br />

• Larger areas of meadow and riparian<br />

habitat<br />

D. Taylor

---

Scale<br />

• understory layers, ground cover (e.g. thick<br />

duff or bare ground)<br />

• vegetation layer- small patches of riparian<br />

vegetation, wet meadow, wet habitats<br />

• forest openings, canopy cover<br />

USFS<br />

USFS<br />

Other Challenges<br />

• substrate within perennial streams, dry<br />

gullies predicted as perennial streams<br />

• species not known from LTBMU- less<br />

information on suitable habitat<br />

• geologic restrictions

---

Detection of 8 new occurrences of 3 sensitive species:<br />

1. Botrychium crenulatum (BOCR2, BOCR3)<br />

(South Lake Tahoe & Zephyr Cove)<br />

2. Draba asterophora v. asterophora (DRASA 1p, 2n, 2o)<br />

(Job’s Sister & Heavenly Ski Area)<br />

3. Lewisia longipetala (LELO 4a, 4b, 4c)<br />

(Jack’s Peak)<br />

USFS<br />

USFS

---

Future Steps<br />

• Equal effort: under-prediction vs. over-prediction<br />

K. Morse<br />

• Functionality of the model vs. true suitable habitat<br />

detection<br />

• Goal: 80% accuracy<br />

• Pinus albicaulis: use TEUI layer? Existing Veg Map?<br />

Remaining points?<br />

• LIDAR & multispectral imagery? Maybe for trees<br />

but not herbaceous species<br />

K. Morse

---

Thanks to: Kurt Teuber, Kyle Christie, Cristina McKernan, Dionne Uzes,<br />

Sean Glantz, Diku Sherpa, Katie Heard, Mike Jennings<br />

USFS