CALIFORNIA STATE UNIVERSITY, NORTHRIDGE

HABITAT SUITABILITY ANALYSIS OF THE COAST HORNED LIZARD

(PHRYNOSOMA CORONATUM BLAINVILLII) IN THE SANTA MONICA

MOUNTAIN NATIONAL RECREATION AREA

A thesis submitted in partial fulfillment of the requirements

For the degree of Master of Arts in

Geography

By

Eryn Morrigan

May 2013

The thesis of Eryn Morrigan is approved:

Shawna Dark, Ph.D. Date

Yifei Sun, Ph.D. Date

James Hayes, Ph.D., Chair Date

California State University, Northridge

! ii! DEDICATIONS

This thesis is dedicated to:

Ryan Morrigan

“For small creatures such as we, the vastness is bearable only through love.”

- Carl Sagan

Exzelia “Zell” Butler

“…The grey rain-curtain turned all to silver glass and was rolled back, and [s]he beheld white shores and beyond them a far green country under a swift sunrise."

- Gandalf

! iii! ACKNOWLEDGMENTS

I would like to first acknowledge my committee chair, Dr. James Hayes, for his guidance and support in this research. I deeply appreciate your patience as my advisor, your feedback on my thesis as well as in the classes I was privileged enough to have with you, and your advice for the future. To my committee members, Dr. Shawna Dark and

Dr. Yifei Sun, thank you for reading my drafts and providing feedback. I am very fortunate to have had your advice and support in this process. To Dr. Soheil Boroushaki, thank you for helping me transform my data into an interactive web-based map as part of my portfolio.

Personally, I would like to thank Christopher Schultz for reading my thesis, but most importantly his indomitable sense of optimism even in the darkest and longest of days. And finally, the Coast Horned Lizard for being such an irresistible and fascinating organism to study. May your species inspire others to respect and protect all the wonders of life on our fragile planet.

! iv! TABLE OF CONTENTS

Signature Page ii Dedications iii Acknowledgment iv List of Figures vii List of Tables viii Abstract ix

1.0 INTRODUCTION 1 1.1 Purpose 2 1.2 Objectives and Research Questions 4

2.0 STUDY AREA 8 2.1 Flora 9 2.2 Fauna 11 2.3 Climate 14

3.0 COAST HORNED LIZARD ECOLOGY AND CONSERVATION 15 3.1 Coast Horned Lizard Distribution and Ecological Suitability 15 3.2 Modern Conservation Concerns for Coast Horned Lizards 18 3.3 Horned Lizards and Prey 20 3.4 Horned Lizard Habitat Issues 23

4.0 METHODOLOGY 27 4.1 Data Sources and Limitations 27 4.2 Mapping Techniques 31 4.3 Classification of Suitability 33 4.3.1 Vegetation Classification 34 4.3.2 Soil Classification 36 4.3.3 Slope and Border Classification 39 4.4 Reclassification 41 4.4.1 Vegetation 41 4.4.2 Soil 46 4.4.3 Slope 51 4.4.4 Border 55 4.5 Landscape Metrics 57

5.0 RESULTS 64 5.1 Expectations 64 5.2 Suitability Models 65 5.2.1 Suitability Model 1 67 5.2.2 Suitability Model 2 71 5.2.3 Suitability Model 3 75 5.3 Metrics 79 5.3.1 Suitability Model 1 Metrics 79

! v! 5.3.2 Suitability Model 2 Metrics 80 5.3.3 Suitability Model 3 Metrics 81

6.0 DISCUSSION 83 6.1 Metrics and Patterns 83 6.2 Research Question 1 86 6.3 Research Question 2 87 6.4 Research Question 3 89

7.0 CONCLUSION 92 7.1 Limitations 92 7.2 Impact on Future Research 94 REFERENCES 96 APPENDIX A 109 APPENDIX B 111 APPENDIX C 112

! vi! LIST OF FIGURES

Figure 1 - Official Map of SMMNRA 8

Figure 2 - SMMNRA Border and Coast Horned Lizard Range 9

Figure 3 - Coast Horned Lizard, Los Angeles County 17

Figure 4 - Juvenile Coast Horned Lizard, Los Angeles County 18

Figure 5 - CHL Habitat Example, San Gabriel Mountains 25

Figure 6 – CHL Habitat Example, San Bernardino Mountains 25

Figure 7 - Original Vegetation Map 44

Figure 8 - Reclassified Vegetation Map 45

Figure 9 - Original Soil Series Map 49

Figure 10 - Reclassified Soil Series Map 50

Figure 11 - Original Slope Percent Map 53

Figure 12 - Reclassified Slope Percent Map 54

Figure 13 - Original SMMNRA Buffer Map 56

Figure 14 - Reclassified SMMNRA Buffer Map 56

Figure 15 - Map of the Four Major Identified HSAs in SMMNRA 66

Figure 16 – Suitability Model 1 70

Figure 17 – Suitability Model 2 74

Figure 18 – Suitability Model 3 78

! vii! LIST OF TABLES

Table 1 - Vegetation Type Classification Table 36

Table 2 - Soil Type Classification Table 37

Table 3 - Slope and Border Type Classification Table 40

Table 4 - Approximate Landmark Boundaries of the Four Identified HSAs 65

Table 5 – Suitability Model 1 Total Area Type Percent 69

Table 6 – Suitability Model 2 Total Area Type Percent 73

Table 7 – Suitability Model 3 Total Area Type Percent 77

Table 8 – Model 1 Metrics 80

Table 9 – Model 2 Metrics 81

Table 10 – Model 3 Metrics 82

Table 11 – Weighting and HSA/HMSA Results Comparison 85

! viii!

ABSTRACT

HABITAT SUITABILITY ANALYSIS OF THE COAST HORNED LIZARD

(PHRYNOSOMA CORONATUM BLAINVILLII) IN THE SANTA MONICA

MOUNTAIN NATIONAL RECREATION AREA

by

Eryn Morrigan

Master of Arts in

Geography

The Coast Horned Lizard (Phrynosoma coronatum blainvillii) is a native species of special concern in California. One of the major threats to Coast Horned Lizards, particularly in southern California, is habitat loss and fragmentation. Studies have suggested that the Coast Horned Lizard selects habitats in chaparral and coastal sage scrub vegetation with friable sandy soils and low to moderate slope percent rise. The

Santa Monica Mountains National Recreation Area (SMMNRA), located in Los Angeles and Ventura counties, is within the known range of Coast Horned Lizard inhabitation.

Due to ecologically disruptive human activity in and around the SMMNRA, potential

Coast Horned Lizard highly suitable habitat may be at risk of fragmentation, decrease in suitability and/or complete loss of habitat. This study identifies areas of high Coast

Horned Lizard suitability so that these areas may be further studied in regards to actual species presence and human impact. Habitat suitability models and landscape pattern were analyzed to investigate the extent to which the SMMNRA is suitable for Coast

! ix! Horned Lizard inhabitation. First, three habitat suitability models were used to classify the habitat suitability of areas in the SMMNRA depending on designated habitat needs such as slope, soil type, and vegetation cover. Then landscape metrics were used to measure habitat fragmentation, particularly for highly suitable areas.

In the SMMNRA, 70% of the total land area was found to meet the majority of the habitat suitability criteria (high suitability, HSA) or some of these requirements

(high-medium suitability, HMSA) for the Coast Horned Lizard. HMSAs covered from

59% to 66% of the SMMNRA, depending on the model’s weighting. Similarities in percent areas were observed between the results for model 1 and model 3. Model 2 showed significantly higher HSA percent (11%) than other models (3%) due to increased vegetation weighting. The landscape patterns and metrics of the SMMNRA showed that most of the area is at least HMSA for Coast Horned Lizards. Fragmentation occurred in areas of human impact but did not prevent the connection of suitable habitats within the interior mountain range. Other factors of suitability should be explored in respect to the

SMMNRA to gain a more accurate assessment of Coast Horned Lizard habitat. In addition, field surveys would be needed to confirm actual Coast Horned Lizard presence

! x! 1.0 - INTRODUCTION

Mediterranean ecosystems have been highly favored and impacted by humans for habitation, agriculture and recreation. These rare and very biodiverse ecosystems compose roughly two percent of the Earth’s land area, are highly altered by human activities, often contain major cities, and have little undisturbed area (Swenson and

Franklin 2000). Southern California has been identified as a Mediterranean ecosystem and a biodiversity hotspot due to the unique flora and fauna present, even though it is highly impacted by humans for development, housing and transportation. The high degree of urbanization along the southern California coastline has resulted in the loss of significant wildlife habitat and the fragmentation of naturally occurring biological corridors (National Park Service 2012 “Climate”).

Habitat fragmentation is the leading factor in species loss both locally and globally (Wilcox and Murphy 1985). It can be caused by increasing the number of habitat patches, decreasing interior habitat area, and increasing the extent of landscape edges or increasing isolation of habitat patches (Li et al. 1993). Birds and mammals are often the first to be affected by habitat fragmentation due to their large ranges, resulting in a decrease in population viability from limited mate selection ranges (Beier 1993).

Reptiles and small mammals with limited mobility may be separated into distinct populations due to narrow geographic barriers such as intervening roads and borders that smaller species cannot bypass (Quinn 1990).

Urban development is an important environmental factor because it creates highly modified landscapes that show higher rates of invasive, non-native, and disturbed

! 1! vegetation (Swenson and Franklin 2000; Endriss et al 2007). Urban modified landscapes often contain small, isolated patches of native vegetation also inhabited by greatly altered flora and fauna. Urbanization affects the physical structure and species composition of native vegetation in many ways, including replacement of native vegetation with invasive or non-native species, soil compaction, changes in microclimates, and isolation of undeveloped native habitat remnants (Endriss et al 2007). Urban development also often involves the increasing encroachment of roads and structures into non-developed land.

One inhabitant of the California Mediterranean ecosystem zone is the Coast

Horned Lizard, Phrynosoma coronatum blainvillii. The Coast Horned Lizard (CHL) is a specialist species in regards to both habitat and prey. Many of the issues the CHL faces in its future conservation are also faced by other specialist species. Invasive species, habitat destruction and fragmentation are all key topics in horned lizard conservation plans

(Brattstrom 1997). Areas of focus in horned lizard research include understanding phylogeographic variation in CHLs, determining suitable habitat areas and the level of prey exclusion by invasive species in known CHL habitats (Fisher, Suarez, and Case

2002). By integrating these three areas, future conservation plans can be built around protecting critical habitat features, prey availability and habitat continuity (Brussard

1991; Kowarik 2011).

1.1 Purpose

Habitat mapping for CHL relies on known suitable habitat characteristics; however knowledge of CHL suitability factors has not been applied to analyze CHL habitat suitability in SMMNRA (CDFG/CIWTG 2007). The California State Fish and

! 2! Wildlife Department has determined the general suitable habitat characteristics of CHLs in previous studies but that information has not been directly used to analyze the Santa

Monica Mountains National Recreation Area (SMMNRA) for habitat conservation effectiveness (Leaché et al 2009). This thesis uses the suitable habitat characteristics of

CHL to determine suitability levels for CHLs in SMMNRA. The results produced by this thesis will fill a gap in research by providing a habitat suitability model for CHLs within

SMMNRA. This study will specifically address identifying highly suitable areas (HSA).

Highly suitable areas will be defined as areas that fulfill all of the suitable habitat characteristics in the California Wildlife Habitat Relationships (CWHR) database, compiled by the California Department of Fish and Game, for their Level II analysis of

CHLs as well as other known habitat needs (Pianka and Parker 1975; Brattstrom 1997;

CDFG/CIWTG 2007).

Fragmentation of HSA by increased human activity (e.g., land development) within SMMNRA could pose serious problems for conservation of CHLs. Fragmentation limits the size of mating populations and increases the likelihood of invasive ant presence due to human activity. Interspersed areas of low suitability are acceptable as long as the percent of acceptable soil type within an area is greater than 75% (CDFG/CIWTG 2007).

This allows for migration through low suitability areas. Although home range migration and long-distance travel is rare, finding new territory is important for CHLs experiencing push factors such as human impact or development, invasive ant spread or wildfires.

High levels of fragmentation in chaparral/CSS vegetation may cause CHLs to be restricted to small, untenable habitats (Brooks et al 2002). Although the known CHL range includes almost all of SMMNRA, it is not known whether that range is uniformly

! 3! suitable for CHL inhabitation.

A number of characteristics that affect the suitability of habitat for CHL (e.g., slope, vegetation type, soil type, and human impact) differ greatly within SMMNRA, suggesting that the range may not be uniformly suitable. This research will analyze variations in soil type, slope, exterior border impact and vegetation and how this variability shapes the spatial pattern of CHL habitat within SMMNRA. With increased human activity in SMMNRA and its resulting disturbance of native habitat, the need for a clear understanding of CHL suitable habitat and its characteristics are essential to future

CHL conservation efforts. This research may also assist biologists in planning field studies of CHLs in situations of limited funding and sample site availability.

1.2 Objectives and Research Questions

Highly areas exist due to specific combinations of environmental conditions that coincide to create suitable CHL habitat. The combination of environmental conditions needed to form an HSA is expected to be spatially biased, only occurring on particular topographic positions, with specific soil conditions, vegetation, slope and border impact.

This thesis asserts that SMMNRA is not a uniformly suitable habitat for CHLs. The suspected reason for this is that highly suitable areas (HSAs) exist due to specific habitat characteristics and variations. The goal of this research is to create a more detailed understanding of CHL habitat suitability using spatial analysis and mapping techniques.

More specifically, three questions will be examined:

! 4! 1. Where are the three largest HSAs?

Conservation planning requires knowing where a species is likely to be and where the most suitable habitat is located. This research provides an initial analysis of where those locations are according to basic CHL habitat needs. Three HSAs will be chosen in order to assess where these HSAs are located in proximity to each other. This may provide insight into potential CHL distribution within the SMMNRA. Although more

HSAs could be identified, this study limits the analysis to three areas so that each area can be more closely examined while still providing variation between areas that can be compared and contrasted. The three largest HSAs will be identified as core habitat zones that could form the foundation of future research, such as population surveys and sampling of CHL populations. Since this research is map-based, the identification of

HSAs will provide useful data to other researchers doing fieldwork as well as conservation planners.

2. Where are HSAs found in relation to park borders?

The distance of HSAs to human impact areas like borders is of particular concern since SMMNRA is within a largely urban area. Human access through borders may be a pressure on CHL habitat, as they actively avoid areas heavily used by humans. If greater distances are found between HSAs and border impact areas, this could indicate that habitat near the borders are more affected by human impact. CHLs may be experiencing greater pressure within the park that limits their expansion into otherwise suitable habitat.

If HSAs are found close to park borders, this could indicate that human impact may be less of a concern at present time. HSAs found near park borders may require even greater

! 5! protections to ensure that they are not impacted in the future by nearby human activity due to their proximity. Future research could use this information to assess potential impact of increased park usage and plan for the conservation of HSAs near border areas.

3. Do HSAs appear to be fragmented in SMMNRA?

Fragmentation of habitat remains a key concern in horned lizard research.

Fragmentation is evidenced by large HSAs broken up by small or narrow areas of less suitable habitat. The term fragmentation can refer to many issues such as habitat loss, change in habitat configuration and patch scale/size/isolation changes (Rose 1982; Fahrig

1997; Martin and Murray 2010). Many factors may cause fragmentation in SMMNRA such as slope differences on a rock face, aquatic features or human impact areas. Such fragmentation may be manageable if it consists of moderately suitable habitat but zero suitability conditions may also be present, creating geographic isolation. Fragmentation can cause horned lizard territories to become nonfunctional for basic behaviors like feeding and mating, leading to a decrease in population (Suarez, Richmond, and Case

2000; Fisher, Suarez, and Case 2002; McIntyre 2003; Endriss et al 2007). Future conservation plans may benefit from further analysis of these areas, if they exist, in order to arrange wildlife corridors for the exchange of mating populations.

These questions help show whether existing HSAs are influenced by human interaction and whether SMMNRA is fragmented into regions of different habitat suitability. This research could lead to future studies as to why this has occurred and into potential solutions including habitat corridors. As CHL habitat suitability is based on

! 6! microhabitat characteristics, future microscale studies could be performed to better understand CHL distribution and habitat use. Future conservation plans may involve increasing hiking paths or access roads through an area that could affect HSAs. This research will provide a baseline survey of CHL habitat suitability guide such future land management decisions.

! 7! 2.0 – STUDY AREA

The Santa Monica Mountains are an east-west mountain range adjacent to the Los

Angeles Metropolitan Area. The SMMNRA administrative boundary encompasses

60,000 hectares. The SMMNRA is bordered on the south by the Pacific Ocean, to the north by suburban communities, to the east by the Los Angeles metropolitan area, and to the west by agricultural lands in Ventura County (Swenson and Franklin 2000; Xu 2001).

It is one of the largest areas of protected Mediterranean ecosystem in the world and contains over 800 kilometers of trails for recreational use. The SMMNRA cuts a wide swath through the western Los Angeles creating a diverse landscape including north and south-facing mountain slopes, coastal areas, valleys, canyons, and multiple vegetation communities. The SMMNRA is generally split into four core regions: Point Mugu State

Park, Malibu Creek State Park, Topanga State Park and the Simi Hills (Swenson and

Franklin 2000; National Park Service 2006).

Gold en

NORTHRIDGE State F re

M e y way

a d o CAMARILLO a o

Ventura Freeway o CHEESEBORO/ w R r e p SEPULVEDA SUN

a To Ventura and PALO COMADO e

r r

boat access to Channel k VALLEY F

CANYONS 101

Islands National Park

o 35mi 56km E J O V A C O N R L L E

Y g

o 5

e

a LA i d S S A N F E R N A N D O V A L L E Y

D

23 V n I yn R L G n Simi Peak NA E I FL a CH AT C N

P S H E CANOGA A THOUSAND OAKS E S L E O S O PARK E C l d K B Civic Center Drive A a n O a N C n C R McCloud Avenue COUNTY VENTURA o R O Y 170 o O

n ad O M e COUNTY ANGELES LOS

jo N 27 National Park Service

Ro A a NEWBURY d C

D Headquarters A Las Posas Road e PARK O v W N

i

H Y r i e ll

c s O

D res ive t r C D t 405 d d OXNARD d l N r a A

a a h a

Borc

k o o N

e R Road R d Y

oa ynn

R L TARZANA

B O

l PLAIN n

v N y

T d o Lewis Road o

d h y n o

Satwiwa Native n i u n e e sa a VAN NUYS H R n C BURBANK American Indian W d o

l O l Culture Center a o y k r Road NATIONAL RECREATION AREA s w e e W st P ero T o d o tr Via Bo o n BOUNDARY u p WOODLAND o

Goleta le i

va L a d rd M Free n w WESTLAKE d ura ay LO CANYON HILLS 101 Vent g V N ADE R u en

R C G G l tur

a a Hueneme Road d h

VILLAGE o d o B v AGOURA ou l v l y ll ev l a o l B n tura a a ENCINO rd d 134 en r V w n d B a d e e a w

HI k d b re o Dri D a n HILLS F ve RANCHO D tl R E s a N N e e a a s s a W o a V t A s O SIERRA VISTA/ L K e a d e Wood Road Wood LE R lab Lake C Y Y h a e C r MARVIN BRAUDE s Encino SATWIWA d a C N 23 e Sherwood 101 n SHERMAN a MULHOLLAND R A Reservoir C d o y N Ag Roa o C a P oura R GATEWAY otrero CALABASAS OAKS n O R n oa WILACRE d y Y Gate PARK C o s N o e M n PARK r u COLDWATER A Lake Ladyface l n B h n o Fryman e O A C e l l v Mountain l R CANYON Eleanor T l g a (gate v 101 R d Road l d I r n E se e Mu F U i d L c L lh r t PARK D N N R L closed t ion a o e oad V o O n R l a d F an o A ve Gate u la e O Y K h ) nd i s v O a B cle r w N l e

O d D

A s A a B l

C a r r

i C rive y

C D 1 W L v Sandstone nd D A e a ll N O SAN VICENTE o n Peak l lh o D Las Virgenes Y PARAMOUNT d u y ri MOUNTAIN M ve O E N n

R N Reservoir RANCH PARK a T O y T C C M o a M e State Park Santa Monica a 27 n Y p i B n E CIRCLE X k Troutdale Calabasas R l FRYMAN A a y N e a w C O l Western U k District n o Road r v CANYON B t Peak Stone n Mountains h r Y RANCH g S n

S s Eagle e S VENTURA CO Town T a D

D e a I U Headquarters r r g C Canyon l RUNYON B i a F

l y LAUREL e c Springs L Laguna Peak A G kb i W Visitor Center C L I o a n

L n r H k

e T RED ROCK L

O C A Reservoir CANYON J B LOS ANGELES CO o

G CANYON PARK

a

(U.S. Navy) A a I

y

L N V B L

CANYON l

Y n FRANKLIN PARK

Mu e n l E lh Y A

LL Little Sycamore y v o TOPANGA n

A POINT MUGU il a PETER Malibu d l PARK O o T V ra l ROCKY N CANYON William O. Douglas T Canyon Road a C e Stu n O n G n

o n Century Lake N b Fuller Avenue Back STRAUSS Lake O R P B d t STATE PARK OAKS o o r Outdoor Classroom A A C

d lhollan a T e RANCH Mu d N u A VALLE H TOPANGA l t Mugu ANO Y B G e Sunset Boulevard R i a Mugu ER gh U KING GILLETTE A N v S T R a Peak R w w in o ay d STATE PARK Y ar a T RANCH r d E d d Lagoon th MALIBU O l S B o o N COLD B e C SPRINGS STUNT RANCH C E a v ific V nt Trippet Ranch WEST c i r e Pa ew CREEK C ada Rd c Pacifi yon Castro MALIBU CREEK Trail k r c C n Road PRESERVE ne A l o a y ast N TAPIA PARK o b HOLLYWOOD C b N Road Peak

H k o D O Y ig l STATE PARK c n d a a O r h Y in B e w R c CASTRO N N a DEER n T POINT MUGU k ARROYO E y A CREST E e CREEK C S M e SEQUIT d a BEL AIR THORNHILL r B T ackb d 27 E Getty CANYON on Saddle Peak R C E e T d r ker ra S BROOME R ec Sch T i le S a r D o l C Museum O d o R n P A i e a Roa l 23 e e P A l BEACH d A a T e o y T S r k N a M I e o R A N u R li L D L h T A U Sc o p s w O Lechusa C d a A a C a Q H a d a d C Sycamore H A o BEVERLY C Road n A Y n E I T R Y e A S P S e S N iu g s N Cove u L m N F u a alib a Y HILLS B CHARMLEE M E E S O BRENTWOOD S d O ZUMA/TRANCAS K C C Rambla L L O C Z

C a a a o L N T n Y WILDERNESS Z n a d r r a n a D b A a L r Pacifico t e

U a C T a n

r a O n s CANYONS i BASS ROCK 1 I S l r d N g e l PARK r e A o M y i R LEO F a L T D l o v v d k o Y

l R Y e C N l R D I e

o u i A o a C Tuna n o CARRILLO TEMESCAL e a B O C u A C d r Y t r

o E a h n g e n e e m a a B l STATE R N O GATEWAY s

y s y o

k C n o l n

E o e v u o N u c C

PARK A y R n d PARK S

S CORRAL

e n

a

o

C

N F

M

n n R

D n R

CANYON O r

o

Y y

o

o

y R e n

o WILL ROGERS

N a

O a

n o L e

R o

n d

d o COUNTY LINE BEACH a D

a w

y

o

N STATE HISTORIC PARK

C s

d R

I a n Getty En SOLSTICE a

cinal D L

a o d i y o LOS ANGELES a C Temescal O CANYON Villa n SEQUIT POINT C Adamson House d e Canyon u A ay s ib hw l Coa ig Road N a Pacific st H NICHOLAS C Y A O M MALIBU COUNTY BEACH N N Malibu Pier EL PESCADOR Y 1 LAS TUNAS PACIFIC PALISADES PARK 10 EL MATADOR O TOPANGA N Moni

TEMESCAL ca STATE BEACH e F r BEACH BEACH nta reeway STATE BEACH iv LAS FLORES Sa r D Malibu Lagoon CANYON PARK

D l l BEACH

DAN MALIBU

a Highway

t

s WILL ROGERS s LA PIEDRA oa BLOCKER BLUFFS 10 n C BEACH o STATE BEACH h c BEACH MALIBU LAGOON 1 c B Pacifi CULVER ZUMA BEACH s STATE BEACH

u

B CITY COUNTY PARK Rainsford SANTA MONICA BEACH SANTA MONICA Place S A N T A PACIFIC OCEAN M O N I C A PALISADES PARK B AY Santa Monica Pier

1 POINT DUME 405 STATE RESERVE POINT DUME Li (WESTWARD BEACH) nc ol n

North B o u le v a VENICE r d Parking Primary road Mulholland Scenic Trail Santa Monica National Park Preserves and Other land within Contact the National Park Corridor (paved) Mountains National Service land private recreation the authorized park Service for information about Campground Recreation Area sites boundary public lands added to the 0 1 Kilometer 4 authorized boundary National Recreation Area since this map was printed. Group Secondary road Mulholland Scenic The park boundary extends to Other public Military land Land outside the 0 1 Mile 4 campground Corridor (unpaved) the mean high tide line along parkland (closed to public) authorized park the coast. boundary MARINA Backcountry DEL REY campground

Figure 1 - Official Map of SMMNRA. Credit: National Park Service

! 8!

Figure 2 - SMMNRA Border and CHL Range. The SMMNRA border is red.

The CHL range boundary is black; present to the north and west (Zeiner et al 1997).

2.1 Flora

The Santa Monica Mountains are the most accessible and largest natural area adjacent to the western Los Angeles Basin and also an important an important area for the diversity of vegetation in southern coastal California. Many factors including fire history, soil diversity, moisture regimes, and topography all combine to create complex and often integrated patterns of woodland, chaparral, coastal scrub and grassland vegetation (Soltz 1984; Swenson and Franklin 2000). All vegetation communities found within the Santa Monica Mountains National Recreation Area are at least minimally hospitable to Coast Horned Lizards, but open areas of low-lying scrub are more suitable ! 9! because they offer more exposure to solar radiation needed for CHL basking requirements.

One of the primary vegetation types in SMMNRA is the coastal sage scrub (CSS) community, a California native, drought deciduous, soft leaf shrub community (Davis et al 1994, Yetter 2012). Common CSS plant species include for example the California sagebrush (Artemisia californica), California buckwheat (Eriogonum fasciculatum), white, purple, and black sage (Salvia apiana, S. leucophylla and S. mellifera, respectively), California encelia (Encelia californica), laurel sumac (Malosma laurina), prickly pear (Opuntia littoralis), lemonadeberry (Rhus integrifolia), brittlebrush (Encelia farinose), Yucca (Yucca whipplei), and coyote brush (Baccharis pilularis) (Davis et al

1994, Yetter 2012). CHLs find CSS and chaparral type vegetation that is prevalent in

SMMNRA to be most suitable.

The area contains several locally common but regionally restricted species such as

Greenbark ceanothus (Ceanothus spinosus), bigpod ceanothus (C. megacarpus), coastal buckwheat (Eriogonum cinereum), and giant coreopsis (Coreopsis gigantean), each because of its high sociability and abundance defines its own suite of vegetation types

(National Park Service 2006). Other alliances defined by California brittlebrush, purple sage, black walnut (Juglans nigra), and lemonadeberry are widespread in southern coastal California but display a concentrated distribution and a broader variation of vegetation associations, creating greater species and habitat diversity (Davis et al 1994;

National Park Service 2006; Yetter 2012).

The Santa Monica Mountains are the westernmost and lowest of the transverse ranges of southern California. The area sustains an extensive presence of some higher-

! 10! elevation chaparral alliances such as hairy ceanothus (Ceanothus oliganthus) and redshanks (Adenostoma sparsifolium) (National Park Service 2006). SMMNRA is also home to the southernmost stand of Valley oak (Quercus lobata) in California and among the largest remaining woodlands of black walnut (Swenson and Franklin 2000). The seaward bases of the mountains have succulent coastal scrub including stands of various prickly pear cacti (Opuntus spp.) along with drought deciduous scrubs including purple sage and California sagebrush (National Park Service 2006). The core of the mountains includes thousands of acres and varied examples of ceanothus alliances (Davis et al

1994). Riparian vegetation includes extensive woodlands that often interface with lower slope woodlands.

2.2 Fauna

The SMMNRA is home to an incredibly diverse spectrum of animal life. This animal diversity is connected closely to the plant life present that forms the ecological foundation of the area (Dratch and Mehrhoff 2002). SMMNRA supports over 450 species of vertebrate life as well as substantial invertebrate populations. This faunal profile is especially unique given the area’s proximity to the heavily urbanized landscape and human population center of Los Angeles (Soltz 1984; Swenson and Franklin 2000).

Amphibians are minimally represented with only ten amphibian species present.

Fish are limited as well, although found in greater diversity and number in the coastal lagoons rather than the inland rivers and streams (Dratch and Mehrhoff 2002). The drier climate and limited water resources preclude widespread amphibian habitation but the more moist riparian areas present within park provide some amphibian habitat (Xu 2001).

! 11! Water pollution, habitat modification by recreation usage, and invasive species present the greatest threats to amphibian populations in SMMNRA (Cunningham 1955).

Amphibian species present in SMMNRA include slender salamanders (Batrachoseps spp.), western toads (Bufo boreus), tree frogs (Hyla spp.) and the federally threatened

California red-legged frog (Rana draytonii). The American bullfrog (Rana catesbeiana), considered an invasive species in competition with other native frogs, has been introduced to the park region. Fish are primarily found in the coastal lagoons, with only a few native species located within the mountain rivers including the federally endangered steelhead trout, as well as locally spawning Pacific lamprey and California grunion

(Dratch and Mehrhoff 2002).

Reptile species represent a large proportion of SMMNRA wildlife, including many unique or threatened species. Although turtles are rarely found due to limited water, there are many commonly observed lizards and snakes (Cunningham 1955).

Besides the Coast Horned Lizard, SMMNRA is also home to, among others, alligator lizards (Elgaria spp.), fence and side-blotched lizards, skinks and the shy legless lizard

(Anniella pulchra). Snakes are well represented with locally common species such as kingsnakes, racers, garter snakes and the only venomous snake found within the park area, the Southern Pacific rattlesnake (Crotalus viridis) (Cunningham 1955). Although the Western diamondback rattlesnake is found within the Los Angeles area, it is the

Southern Pacific rattlesnake that is found commonly in the SMMNRA range. This presents a recreational danger within the area of rattlesnake bites, although hospital treatment is within close reach due to SMMNRA’s proximity to urban areas.

Birds are a critical part of the SMMNRA ecosystem and more than 380 species

! 12! can be found year-round. The Los Angeles area, SMMNRA specifically, acts an important stop along the Pacific Flyaway (Soltz 1984). This north-south migration route is essential to the survival of many species. Only 1/3 of all species within SMMNRA actually live and breed in the area; the rest are temporary or seasonal residents due to the

Flyaway location (Yetter 2012). The Mediterranean climate provides a temperate transition zone that reliably produces food from migration journeys (Soltz 1984). Some of the bird variety includes waterfowl, sparrows, finches, owls, raptors, wrens and hummingbirds. Golden eagles (Aquila chrysaetos) are also present in grassland portions of SMMNRA (Plantrich 1990). There are hundreds of bird species that act as essential links within the ecosystem. Bird watching is a common activity for humans within the park and monthly “birdwalks” are held to encourage both species identification and human appreciation of the avian diversity.

Mammal presence within the Santa Monica Mountains National Recreation Area is well documented and ranges from very small rodents all the way to large carnivores such as mountain lions (Quinn 1990; National Park Service 2012). Many species are seed-dispersing rodents that humans may consider pest species but are in fact, critical aspects of the ecosystem. Avian raptors, snakes and larger mammals rely on the park’s rodent populations as an important food source (Quinn 1990). Fragmentation of habitat within SMMNRA and the adjoining urban areas is a significant issue for many mammals that live there but specifically mountain lions and bobcats (Swenson and Franklin 2000;

National Park Service 2012). Park officials have many projects to track several mammal species within the park (Dratch and Mehrhoff 2002). .

! 13! 2.3 Climate

The Santa Monica Mountains National Recreation Area is often referred to as

“chaparral” or Mediterranean ecosystem. These terms are sometimes also used to describe its climate, as the Mediterranean ecosystem has distinct climate features.

Mediterranean climates occur only in five small zones around the planet, found between roughly 30 and 40 degrees latitude (north and south) and they are located along the western edges of continents. Cold ocean currents moderate Mediterranean climates, which are characterized by mild, rainy winters and warm, dry summers (Bailey 1966;

Cody 1986). Most precipitation in southern California occurs between November and

April but varies depending on elevation and distance from the ocean (Huffman 1998). In

North America, the Mediterranean climate zones extend intermittently along the west coast of the United State into northern Baja California. Other Mediterranean areas include the central Chilean coast, southwest Australia, the Cape region of South Africa, and eponymous Mediterranean basin.

! 14! 3.0 – COAST HORNED LIZARD ECOLOGY AND CONSERVATION

3.1 Coast Horned Lizard Distribution and Ecological Suitability

The Coast Horned Lizard (Phrynosoma coronatum) includes six unofficial subspecies, the most studied of which is the Blainville’s CHL (Phrynosoma coronatum blainvillii). Recent research suggests that the Blainville’s Horned Lizard is part of the P. coronatum species complex (Leaché et al 2009). A species complex is a group of closely related species, where the exact demarcation between species is often unclear due to their recent and usually still incomplete reproductive isolation. The P. coronatum species complex stretches from central California into northern Baja California (Reeder and

Wiens 1996).

The CHL is a species of special concern (SSC) in California and as such, is protected in much of its range (Fisher, Suarez, and Case 2002). According to the

California Department of Fish and Wildlife’s “Species of Special Concern” publication

(Comrack et al 2008), a SSC is a species, subspecies or distinct population of an animal native to California that currently satisfies one or more of the following (not necessarily mutually exclusive) criteria:

• is extirpated from the State or, in the case of birds, in its primary seasonal or

breeding role;

• is listed as Federally-, but not State-, threatened or endangered; meets the

State definition of threatened or endangered but has not formally been listed;

! 15! • is experiencing, or formerly experienced, serious (noncyclical) population

declines or range retractions (not reversed) that, if continued or resumed,

could qualify it for State threatened or endangered status;

• has naturally small populations exhibiting high susceptibility to risk from any

factor(s), that if realized, could lead to declines that would qualify it for State

threatened or endangered status.

It is not considered a threatened species but in certain parts of its range, it has experienced significant habitat destruction and human impact. California prohibits collecting CHLs without a permit due to their protected status (Fisher, Suarez, and Case

2002). Habitat destruction and history of over collecting by the pet trade present the greatest threat to this species (Price 1990; Jennings 1987). Invasive ants and significant human expansion into native CHL territory within recent years have compromised CHL ecology by modify habitat, displacing prey, and isolating CHL populations (Sherbrooke

2003; Brattstrom 1997; MacMahon et al 2000; Holway and Suarez 2006).

Coast horned lizards select wide-open spaces with coastal sage scrub, sandy washes and hill/mountain regions that are also the most suitable habitat of its main prey, harvester ants (Leaché et al 2009). Pogonomyrmex and Messor ants (all considered harvester ants) are the most common food source for horned lizards but they are capable of significant retaliation against predators (Pianka and Parker 1975). Horned lizard tongues are highly vascularized and sticky to aid in the quick, snapping motions needed to pick up individual ants (Sherbrooke 2003). Invasive ant species competitively exclude native harvester ants, leaving little to no food for CHL populations in the affected areas

! 16! (Sherbrooke 2003; Suarez, Richmond, and Case 2000). A diet of only invasive ant species leads to starvation of CHL and subsequent die-off (Leaché et al 2009).

Habitat disruption remains the key habitat issue, particularly in urban and/or invasive species affected areas (Endriss et al 2007; Martin and Murray 2010). This habitat disruption is often connected with human repurposing of land, particularly for farming, but also human housing developments (Kowarik 2011; Olden 2006; Shearer et al 2006). Urban development specifically causes highly modified patches of habitat and increased horned lizard mortality from insecticides, landscaping activities, and introduced pests such as cats (Endriss et al 2007; Leaché et al 2009).

Figure 3 - Coast Horned Lizard, Los Angeles County. Photo Credit: Gary Nafis

! 17!

Figure 4 - Juvenile Coast Horned Lizard, Los Angeles County. Photo Credit: Gary Nafis

3.2 Modern Conservation Concerns for Coast Horned Lizards

There are two main issues at the heart of CHL conservation. First, the interconnectedness of horned lizards and harvester ants must be understood and considered in any conservation plan. Second, habitat conservation plans need to account for differences in phylogeographic horned lizard groups and be adaptable for both urban and open-land areas so that plans can be adjusted to specific locations (Shearer et al

2006; Syphard et al 2011).

Coast horned lizards cannot be protected if harvester ants are not also assessed and present in protected areas. Conservation plans aimed at CHL conservation must also include measures to protect native ant populations and avoid introduction of invasive ant species. A central issue in wildlife conservation is whether focusing on a single species is more effective than an ecologically linked grouping of organisms (Syphard et al 2011;

! 18! Thomas 2000; Andelman 2011). Protection of multiple species is especially critical when considering the interconnectedness of CHLs and harvester ants. One of the methods to approach modern conservation is for some research scientists to lead large-scale conservation projects rather than do small studies for academic journals (Andelman

2011). By putting research experts in charge of large projects, their expertise can be used directly to design conservation plans. Closer coordination and communication between

CHL and ant researchers could lead to more comprehensive conservation plans based on a holistic view of lizard-ant interaction.

Deciphering CHL research and constructing conservation plans can be difficult because some studies have treated CHLs as different subspecies while other studies refer to all CHL as P. coronatum (Brattstrom 1997; Reeder and Wiens 1996). Genetic analysis concluded that there are no real subspecies of P. coronatum but rather distinct phylogeographic species complexes (Leaché et al 2009; Hodges and Zamudio 2004). A species complex occurs when a group of closely related species cannot be reliably delineated into unique species or subspecies due to recent, often rapid, divergence involving reproductive isolation (Leaché et al 2009). Subspecies are members known to be of the same species but with distinguishably different features, capable of interbreeding and producing fertile offspring, but they generally do not interbreed due to isolations barriers. The P. coronatum phylogeographic species complex is in the process of divergent speciation and although genetically similar, it is not known whether they would all be capable of interbreeding (Leaché et al 2009). Some phylogeographic groups are found in more urban areas, such as the northern and southern California groups, while others like the Baja California group inhabit much more rural and minimally disturbed

! 19! territories. All of these variant phylogeographic groups should be considered unique for conservation purposes since they inhabit specific areas with different conservation concerns (Leaché et al 2009). Distinguishing different species and sub-species is a notoriously difficult task that complicates conservation research (Goldstein et al. 2005;

Hodges and Zamudio 2004). Common concerns though are human-caused destruction of coastal sage scrub and chaparral vegetation, and invasive ant dispersal.

One of the issues complicating CHL conservation is the phylogeographic difference among CHLs along with their respective habitat concerns. The conservation plan for one phylogeographic group will likely differ from another based on urban density, location and habitat type, available conservation land, and the presence of invasive ant species (Reeder and Wiens 1996). Only by taking these many factors into consideration can an effective conservation plan be constructed that ensures the continued survival of the entire P. coronatum species complex (Shearer et al 2006).

3.3 Horned Lizards and Prey

Horned lizards rely on harvester ants (Pogonomyrmex spp.) as their primary food source (Montanucci 1989; Suarez, Richmond, and Case 2000; Pianka and Parker 1975).

The CHL, more than many other horned lizards, faces serious prey depletion due to invasive ant species, e.g., the Argentine Ant (Linepithema humile) and the Red Imported

Fire Ant (RIFA, Solenopsis invicta), that displace and discourage native ant colonies

(Suarez, Richmond, and Case 2000; Leaché et al 2009). Changes in distribution and composition of ant colonies can have a severe impact on horned lizard populations

(Whitford and Bryant 1979; Whiting, Dixon, and Murray 1993). Argentine ants are

! 20! generally not found in horned lizard stomach contents, indicating the inedibility or extreme distaste of Argentine ants to horned lizards (Suarez and Case 2002). Also, juvenile growth rates of CHL are significantly suppressed in the presence of Argentine ants due to lack of prey (Suarez and Case 2002).

The primary issue with prey remains the competitive exclusion of native ants caused by the invasive Argentine Ant and RIFA (Suarez and Case 2002; Sherbrooke

2003). These non-native species can completely eliminate native colonies and render entire areas inhospitable to horned lizards, although they are limited in their invasive range by their intolerance to highly xeric conditions (Suarez, Richmond, and Case 2000,

Moody et al 1991; Wiernasz and Cole 1995). The relationship between horned lizards and their preferred prey (Pogonomyrmex) must be paramount in assessing suitable habitats and predictions of future species survival (Meyers et al 2006).

Appropriate habitat availability limits the distribution of CHL prey, which in turn is a limiting factor for CHL distribution (Powell et al 1998; Leaché et al 2009). Even with high levels of predation by specialist horned lizards, harvester ant colonies are found to be relatively stable and do not suffer from colony depression (Wiernasz and Cole 1995).

Argentine Ants and RIFA lack native predators in their invaded habitats, allowing them to reproduce at a far greater rate and become better established in new habitats where

CHL occur (Pianka and Parker 1975; Suarez and Case 2002; Meyers et al 2006). This exacerbates the initial invasive problem since few other species eat any kind of ants

(Suarez and Case 2002). Although some adaptation to other food sources does occur, this is a disruptive and fragmenting force on horned lizard habitat availability (Whitford and

Bryant 1979; Whiting, Dixon, and Case 1993).

! 21! The location selection process of harvester ant mating swarms largely determines where horned lizards will congregate because sustainable horned lizard populations depend on thriving harvester ant colonies as a part of CHL habitat. Mating swarms initially show a closely clumped spatial pattern but quickly become uniform through competition, although new colonies located farther away from the original have a greater survival rate due to lower competition density (Wiernasz and Cole 1995). This increased distance between successful harvester ant colonies can create a patchy distribution of horned lizards, making population and distribution analysis difficult (Henke 2003;

MacMahon et al 2000; Holway and Suarez 2006). The interaction between native and invasive ants is a determining factor for both horned lizard and harvester ant species survival and to a large extent the presence of healthy harvester ant colonies signifies a high likelihood of successful horned lizard territory (Price 1990; Suarez, Richmond, and

Case 2000; Suarez and Case 2002).

Horned lizards are morphologically adapted to hunting ants and have simple strategies like waiting near anthills to catch them (Newbold, Scott, and MacMahon 2009;

Whitford and Bryant 1979; Montanucci 1989). In general, horned lizards do not exert much energy in hunting ants (Sherbrooke 2003, Pianka and Parker 1975, Lahti and Beck

2008). Ant percentage in the diet shows very significant negative correlations with certain mouth features like the mandible length and mouth diameter while positively correlated with length of the tooth row (Montanucci 1989). Although it may seem that long tooth rows could aid in biting down on ants, nearly all ants are found whole and unbitten in horned lizard stomachs, having been drowned and incapacitated by copious amounts of sticky mucus: the teeth are for defense, not feeding (Sherbrooke 2003). This

! 22! mucus prevents ants from stinging the horned lizard while in the stomach (Sherbrooke

2003). The horned lizard is protected from harvester ant stings and bites by its tough plate armor like scales and specifically, the Texas Horned Lizard has been found to have a factor in its blood plasma that detoxifies ant venom (Sherbrooke 2003). Horned lizards also prefer individually foraging ants, rather than column foraging ants, allowing for easier hunting of prey (Munger 1984 “Long-Term Yield”; Sherbrooke 2001, Whitford and Bryant 1979, Pianka and Parker 1975). Not all horned lizard species chemically taste their food prior to consuming it by flicking their tongue to pick up scents. Tongue flicking appears to be absent in the Texas Horned Lizard and may be absent or diminished in other Phrynosoma species but that remains a question for future research

(Cooper and Sherbrooke 2009).

3.4 Horned Lizard Habitat Issues

Horned lizards select sandy and gravely surface substrate to aid in their digging behaviors. Most horned lizard habitats have areas of loose sand for digging, camouflage, prey availability, and thermoregulation but still retain enough solid footing for movement

(Whiting, Dixon, and Murray 1993; Beauchamp et al 1998). Substrate cohesiveness is an important factor in creating successful burrows, as the sand or dirt must be sufficiently friable to dig in but not too loose or it will collapse (Mathies and Martin 2008; Wone and

Beauchamp 2003). Horned lizards are very well adapted to these environments through thick scales to protect against the sand and gravel, patchy camouflage to match the ground and horns/bumps to disguise their profile (Pianka and Parker 1975; Young et al

2004). They congregate in open, sunny locations for thermoregulation, mating and

! 23! feeding (Sherbrooke 2003; Burrow et al 2001; Lynn 1965; Goldberg 1983). Although most home ranges are fairly exclusive (5 – 25 meters2), there are overlapping territories and minimal male defense of those areas (Fair and Henke 1999; Rose 1982; Munger 1984

“Home Ranges”). Most of their morning behavior occurs in these open areas, particularly basking, but in the afternoon, horned lizards retreat to more protected shrub areas

(Sherbrooke 2003; Pianka and Parker 1975; Burrow et al 2001; Lynn 1965). In areas of very dense vegetation, horned lizards have limited access to basking sunlight and find it difficult to move through the brush due to their squat and spiny bodies but they do use areas of moderate vegetation as protection from predators (Fair and Henke 1997; Munger

1984 “Home Ranges”; Burrow et al 2001; Rose 1982). In the afternoons, scrub vegetation is used as a thermal refuge until temperatures cool in early evening, when activity begins again (Burrow et al 2001; Sherbrooke 2003). Harvester ants also make colonies in open, sandy areas, leading to the horned lizard suitability for such habitats

(Fisher, Suarez, and Case 2002; Whiting, Dixon, and Murray 1993; Burrow et al 2001).

The dearth of vegetation in these areas aids in horned lizards pinpointing prey as well as making the approach of potential predators like coyotes and hawks more noticeable

(Munger 1984 “Home Ranges”; Beauchamp et al 1998). Also, CHL use defensive blood squirting from their eyes when approached by canine predators in open areas, although their primary defensive is running to nearby scrub cover for camouflage (Middendorf et al 2001; Middendorf and Sherbrooke 1992; Sherbrooke and Middendorf 2001;

Sherbrooke 2004; Sherbrooke 2008).

! 24!

Figure 5 - CHL Habitat Example, San Gabriel Mountains.

This has similar CHL habitat characteristics to SMMNRA.

Figure 6 – CHL Habitat Example, San Bernardino Mountains.

Note the barren circle around the harvester anthill in the foreground. This is an indication of an active harvester ant colony. It is also a highly suitable location for CHL basking, mating and feeding behaviors (Goldberg 1983).

! 25! Urban development poses a major challenge to CHL conservation efforts. As development in prime habitat areas continues, availability of lands for conservation is limited. Habitat fragmentation due to urban development, farming, and invasive ants has lead to areas of limited population and even complete exclusion from once dense zones of horned lizard habitation (McIntyre 2003; Endriss et al 2007; Fisher, Suarez, and Case

2002). Although horned lizards can adapt to modified environments, certain aspects such as prey availability, open areas for basking and mating, and proper temperature ranges are necessary for continued habitation by significant horned lizard populations

(Beauchamp et al 1998; Burrow et al 2001). When discussing fragmentation in relation to conservation, it is important to realize that this term can refer to many issues such as habitat loss, change in habitat configuration and patch scale/size/isolation changes

(Fahrig 1997; Rose 1982; Martin and Murray 2010). In areas of fragmentation or modification, horned lizard territory can become nonfunctional for basic behaviors like feeding and mating (Fisher, Suarez, and Case 2002). This can cause a drop in both adult and juvenile survival rates, particularly when combined with competitive exclusion of harvester ants by invasive species (McIntyre 2003; Endriss et al 2007; Suarez, Richmond, and Case 2000). In areas of significant fragmentation, access to females and mating territories can be severely restricted, leading to decreased horned lizard populations

(Donaldson, Price, and Morse 1994; Stark, Fox, and Leslie Jr. 2005; Goldberg 1983).

Also, native ant populations are strongest in areas of the least fragmentation and human encroachment (Suarez and Case 2002).

! 26! 4.0 – METHODOLOGY

This research mapped environmental variables associated with CHL habitation in the SMMNRA and developed from those variables three habitat suitability maps. It also used landscape metrics to measure fragmentation of CHL suitable habitat areas in the

SMMNRA. The habitat suitability maps and related metrics will improve understanding of CHL habitat in SMMNRA and will be useful for CHL conservation efforts. As field studies of horned lizards can be very difficult, it is important to narrow down their most likely habitat locations before any actual field studies are planned (Grant and Doherty

2007). Improved understanding of CHL habitat distribution would help land managers target their efforts make CHL conservation activities and microsite habitat characteristic studies more efficient and successful.

4.1 Data Sources and Limitations

According to the California Wildlife Habitat Relationships (CWHR) database, compiled by the California Department of Fish and Game, there are known habitat suitability parameters for the Coast Horned Lizard under their Level II analysis

(CDFG/CIWTG 2007). These data were the basis for most of the suitability analysis performed in this research. Appendix A lists these specific parameters in greater detail.

These parameters were necessary for creating the final classifications in the models, as well as the percent weights used for those models. The vegetation parameters were already provided in order of suitability as well as soil factors such as friability. These suitability parameters were used to sort the vegetation and soil data into suitability

! 27! categories for the final models. The determination of acceptable, functional HSA patch size was based on minimum patch size and “Spatial Habitat Requirements for Persistence of Population” minimum habitat size data.

Slope data were necessary to model habitable areas for CHL since slope is a limiting factor for CHL habitat suitability. Slope rise greater than 50 percent is not minimally suitable for CHL and were, as such, deemed non-inhabitable. Slope data obtained from USGS DEM (10 m resolution) were used to create the slope percent layer

(USGS National Map). Slope data came from DEM files for SMMNRA region supplied by the United States Geological Service (USGS) and is, of course, limited by scale. One limitation important to this habitat suitability analysis was the dependency on scale

(extent and grain). The scale of a study affects patterns found in the landscape (Wiens

1989; Rahman et al 2003). Both the organism and the scale at which the landscape patterns were apparent influenced the selection of the appropriate scale for this study

(Mayer and Cameron 2003; Rahman et al 2003; Hernandez et al 2006). If the resolution is coarser than the level at which the species perceives the landscape, this can obscure important habitat relationships between the species and their environment (Mayer and

Cameron 2003). Although finer resolution data would have provided a better scale for

CHL habitat needs, the sheer size of the SMMNRA and the resolution of its related available data limited the scale of this study.

The California Department of Fish and Game Biogeographic Data Branch and the

National Park Service provided the vegetation data layers. Vegetation descriptive zones used in this analysis were not highly specific (“super generalized”) and were only meant to give a general idea of distinctions made within SMMNRA (Dennison and Roberts

! 28! 2003; Hernandez et al 2006). Most of SMMNRA is within a minimally acceptable vegetation profile but there are pockets of more suitable types. Soil types were sorted by those characteristics and processed into larger classification groups for the final models.

Vegetation groups were classified using the CWHR analysis data to create to final models (CDFG/CIWTG 2007). These were already presented in preferential order so no processing (sorting) of that data was necessary, only placement within suitability groupings. Vegetation can actually vary greatly on the ground at the horned lizard environmental scale and although the data stated that one vegetation type was present in a given location, the actual ground conditions could have been different due to microscale variation, modification since the data were gathered or remote sensing classification issues (Roberts et al 1998; Rahman et al 2003). Those potential differences were not accounted for in this research. The vegetation classifications were compiled from the

National Park Service’s “Vegetation Classification of the Santa Monica Mountains

National Recreation Area and Environs in Ventura and Los Angeles Counties,

California” published in 2006, which was described as follows:

The U.S. Geological Survey (USGS) and National Park Service (NPS) formed a partnership in 1994 to map the vegetation of the United States National Park system units using The Nature Conservancy's National Vegetation Classification, a standard for reporting vegetation information among federal agencies (Grossman et al. 1998). Goals of the projects include providing baseline ecological information to resource managers in the parks; putting the data into regional and national contexts; and providing opportunities for future inventory, monitoring, and research activities. Each park developing a vegetation map followed a standardized field sampling and vegetation classification protocol to document the various vegetation types found in that park. This information is used by photointerpreters to delineate polygons of vegetation communities, which are subsequently subjected to an accuracy assessment process (USGS 1994). The final products consist of a vegetation map, descriptions of each vegetation type, a key to each type, and all related data and metadata files (original field forms, plot database, accuracy assessment points, etc.). This

! 29! report presents the work at the Santa Monica Mountains National Recreation Area (NRA) (park code: SMMNRA) and environs conducted from 2001 to 2005.

The soil survey used as the basis of soil type analysis and classification in this research was from a publication of the National Cooperative Soil Survey, a joint effort of the United States Department of Agriculture and other Federal agencies, State agencies including the Agricultural Experiment Stations, and local agencies. Major fieldwork for this soil survey was completed in 2001. The Natural Resources Conservation Service, the

United States Department of the Interior and the National Park Service created this survey cooperatively. The survey was part of the technical assistance furnished to the

Ventura County and Santa Monica Mountains Resource Conservation Districts. (United

States Department of Agriculture, Natural Resources Conservation Service 2006.) Soil type is highly variable in general, but specifically within SMMNRA due to differential erosion patterns, rain shadow, drainage watersheds and wind flow. The soil data was from the most recent United States Department of Agriculture’s (USDA) Natural

Resources Conservation Service soil survey of SMMNRA. Only soil conditions within the top six to eight inches of soil were relevant to CHL research since they only dig and burrow within those top friable layers (Whiting, Dixon, and Murray 1993; Beauchamp et al 1998). Friability was the primary determinant for horned lizard soil suitability, according to the CWHR Level II analysis (CDFG/CIWTG 2007). Burrowing is a necessity for all horned lizards for protection and hibernation. Even though soil type was an important factor in CHL habitat selection, the actual soil selection criteria were more detailed, including soil moisture level, grain size, bare rock exposure and soil coloration

(Wone and Beauchamp 2003). Comparison of the CWHR analysis characteristics to the

! 30! soil survey types formed the soil suitability classifications. Matches for friability, gravel, sand, similarity to CHL coloration, and low water content were deemed higher suitability than other soil characteristics (CDFG/CIWTG 2007). Soil types were sorted by those characteristics and processed into larger classification groups for the final models.

Human impact areas (external border only) were determined from the National

Park Service maps and shapefiles of SMMNRA (Clarke 2002). No processing of these files was needed for use in this study, only the formation of a buffer around the provided border.

4.2 Mapping Techniques

The habitat suitability model mapped locations where limiting environmental variables for CHL habitat occurred and created weighted overlays (models) based on known suitability characteristics of CHL (CDFG/CIWTG 2007). These models indicated varying areas of habitat suitability based on the weighted variables of vegetation, soil, slope percent, and border. The habitat suitability analysis was modeled using ArcGIS

10.1 software. This research used the ModelBuilder application in ArcGIS to build the appropriate models for the habitat suitability analysis (ESRI 2009).

Reclassification and weighted overlay techniques were used to create the final habitat suitability maps (Swenson and Franklin 2000). The reclassification technique changes the original cell values for each factor, such as vegetation or soil type, to new classified values representing suitability levels (Swenson and Franklin 2000; ESRI 2009).

The original cell values represented either quantitative or qualitative attributes of a certain feature (e.g., slope, vegetation type, etc.). During reclassification, the variables

! 31! were assigned values of suitability (six values) according to the species habitat characteristics described further in section 4.3. These characteristics and their suitability categorization (reclassification) were derived from the CWHR analysis, soil survey and vegetation data. A common scale (zero to five) of suitability was used in this study. The weighted overlay technique overlaid several of the habitat suitability variables using a common scale and weighted each variable according to its importance (Swenson and

Franklin 2000; ESRI 2009). The weighted overlay technique first enabled the reclassification of the variables, then weighted the variables and combined the weighted variables to create the final habitat suitability index and map.

For example, the soil type procedure was to take the initial values that were soil series types and sort them by the CWHR suitability needs for friability, gravel and coloration into six categories from most suitable to non-suitable. This process of assigning initial soil values into suitability groups used the classification system described in section 4.3.2. The resulting six suitability groups were considered the classified values to be used in the weighted overlays (models). These reclassified groups were required to be consistent through all data types so they could be accurately compared in the final models. These reclassified values ranged from zero to five with five as the highest suitability and zero representing non-habitable for CHLs.

This process was repeated for each data type (soil, vegetation, slope percent, and border) based on the classifications described in the following section, 4.3. After each data type was in a comparable zero to five scheme, three weighted overlays were created.

The weighting was based on the CWHR analysis priorities as well as other studies that had examined the importance of each of those factors in horned lizard habitat suitability.

! 32! Not all of the studies were restricted to CHLs but other horned lizards share the same approximate valuing of habitat characteristics and requirements (Pianka and Parker 1975;

Whiting et al 1993; Brattstrom 1997; Beauchamp et al 1998; Burrow et al 2001; Fisher,

Suarez and Case 2002; Sherbrooke 2003; Endriss et al 2007; CDFG/CIWTG 2007;

Leaché et al 2009).

4.3 Classification of Suitability

The habitat factors that were assessed were vegetation type, soil type, slope percent and border. Final assessment categories rated habitat factors at specific locations from 0 to 5, from zero suitability areas (ZSA) that is non-inhabitable, to high suitability areas (HSA). There were six categories of suitability (zero, low, low-medium, medium, high-medium, and high) each habitat factor. The six final assessment categories were nominally: zero (ZSA), low (LSA), low-medium (LMSA), medium (MSA), high-medium

(HMSA), and high suitability areas (HSA). Areas considered ZSA included slope above

50 percent rise, any vegetation that was aquatic or wetland type, and aquatic soil types

(Pianka and Parker 1975). Human impact did not preclude CHL presence but did act as a negative habitat pressure, represented by the MSA designation of the border buffer area for weighting (Beauchamp et al 1998; Fisher, Suarez and Case 2002; Endriss et al 2007).

These final categories corresponded to probabilities of CHL occurrence; once located through the models, they could be used to identify potentially significant areas of CHL activity and restriction of habitat usage.

! 33! 4.3.1 Vegetation Classification

Vegetation classifications followed a 0-5 scale, with 0 being ZSA, 1 LSA, 2

LMSA, 3 MSA, 4 HMSA and 5 HSA. All vegetation/land use type categories were taken from the National Park Service’s 2006 publication “Vegetation Classification of the

Santa Monica Mountains National Recreation Area and Environs in Ventura and Los

Angeles Counties, California” (National Park Service 2006) Water and wetland types were given a ZSA (0) rating since CHLs do not swim or use aquatic areas (Pianka and

Parker 1975).

Agriculture, urban/disturbed, built up, and cleared land areas were considered

LSA (1) due to their low prevalence of scrub vegetation and native plants (Martin and

Murray 2010). CHLs largely avoid human impacted areas and have lower survival rates in these areas due to habitat destruction. Scrub/shrub cover is necessary for thermal management and prey avoidance but these areas are mostly devoid of this vegetation type.

The LMSA (2) category included rocky outcrops and sandy/rocky/mud types.

Although those appeared to be soil type designations, in terms of vegetation they indicated little vegetation cover and mostly bare rock or soil. CHLs do use these areas occasionally but cannot live solely within these areas due to a lack of vegetative cover and prey food sources (Brattstrom 1997; Fisher, Suarez and Case 2002; Sherbrooke

2003). These areas were small and limited largely by slope within SMMNRA. Since this vegetation type was not prevalent, few areas were given this designation.

Disturbed, exotic and invasive vegetation were classified as MSA (3) since, although not native, certain invasive and exotic scrub species do occur in SMMNRA and

! 34! could provide acceptable vegetation coverage (Brattstrom 1997; Beauchamp et al 1998;

Burrow et al 2001; CDFG/CIWTG 2007; Martin and Murray 2010). Harvester ants do not prefer exotic plant species as food however so this is given a medium designation based on a mix of acceptable cover but lower desirability of prey food sources (Suarez,

Richmond and Case 2000).

The HMSA (4) category contained prairie/meadows and riparian/upland tree areas. This vegetation profile often contained areas of scrub/shrub ground cover, gravelly open patches and some canopy openings for solar penetration, all of which are needed for

CHL habitation (Pianka and Parker 1975; Brattstrom 1997; Beauchamp et al 1998;

Fisher, Suarez and Case 2002; CDFG/CIWTG 2007). These were high-medium suitability areas but if the canopy cover was too thick, solar irradiance could be too low to sustain CHLs. In this research, these were grouped together without specification of canopy cover percentage. Future studies could examine the impact of canopy cover percentage within tree-dominant areas on classification (Dennison and Roberts 2003).

For the purposes of this analysis though, prairie/meadows and riparian/upland tree areas provided enough proper vegetation cover to score as a high-medium suitability area.

Finally, the HSA (5) classification was defined by the two vegetation types considered ideal habitat for CHLs: coastal sage scrub (CSS) communities and chaparral.

In these areas, the vegetative ground cover percentage, mix of native species, and suitable prey food sources made them the highest suitability areas of CHL (Brattstrom 1997;

CDFG/CIWTG 2007). These areas ideally were minimally disturbed, mostly native and match with patches of ideal soil type as well. CHL are known to inhabit chaparral and

CSS zones as their primary choice, even when faced with invasive ant concerns (Suarez,

! 35! Richmond and Case 2000). The final weighting values that gave vegetation a higher percentage in the weighted overlay models indicated the importance of vegetation type to

CHL habitat selection (Beauchamp et al 1998; Burrow et al 2001; Fisher, Suarez and

Case 2002; CDFG/CIWTG 2007).

Table 1 - Vegetation Type Classification Table

Abbreviation Classification Value Description ZSA Zero Suitability Area 0 Water and wetlands.

LSA Agriculture, urban/disturbed, Low Suitability Area 1 cleared or built-up.

LMSA Low-Medium Suitability Rock outcrop, 2 Area sandy/rocky/mud.

MSA Disturbed vegetation, exotic or Medium Suitability Area 3 invasive species.

HMSA High-Medium Suitability Prairie/meadow, riparian and 4 Area upland trees.

HSA Chaparral and coastal sage High Suitability Area 5 scrub (CSS)

4.3.2 Soil Classification

Soil classification was challenging considering the 68 different soil series recorded in the soil survey for SMMNRA (USDA/NRCS 2006). Soil properties of each series were examined for suitability characteristics. For the purposes of this study, the following categories were automatically classed as ZSA: water, dams, pits/dumps and frequently flooded tidal flats (Sulfic Fluvaquents) (Pianka and Parker 1975;

CDFG/CIWTG 2007). All other soil series had at least a minimum likelihood of CHL being present for part of their life cycle. Texture was a key aspect of CHL soil suitability with gravelly/sandy loam and clay preferred (Wone and Beauchamp 2003). Certain soil

! 36! textural features such as non-friability and slow water infiltration rates are not conducive to CHL needs (CDFG/CIWTG 2007). Specific soil series that rated high in CHL suitability characteristics include the Calcic Haploxerept-Mollic Haploxeralf association,

Zumaridge-Sumiwawa complex, and Chumash-Boades-Malibu association. These soils had high friability, at least 15 percent gravel composition, gravelly silt loam, light gray to medium brown coloration, and are generally dry (Beauchamp et al 1998; Fisher, Suarez and Case 2002; USDA/NRCS 2006; CDFG/CIWTG 2007).

Soil habitat assessment categories (Table 2) were based on known suitable soil characteristics and the soil profiles provided in the soil survey documentation (Whiting,

Dixon, and Murray 1993; Beauchamp et al 1998; Wone and Beauchamp 2003; Pianka and Parker 1975; Fisher, Suarez, and Case 2002; Burrow et al 2001; USDA/NRCS

2006.).

Table 2 - Soil Type Classification Table

ZSA (0) LSA (1) • Water • Abaft-Beaches-Urban land complex, 0 • Dams to 5 percent slopes • Pits/Dumps • Abaft-Beaches association, 0 to 5 • Sulfic Fluvaquents percent slopes • Area not digitized. • Botella loam, 2 to 9 percent slopes • Camarillo loam, coastal, 0 to 2 percent slopes • Cotharin clay loam, 30 to 75 percent slopes • Cropley association, 2 to 15 percent slopes • Cropley clay, 0 to 2 percent slopes • Cropley clay, 2 to 9 percent slopes • Cropley, coastal-Urban land- Haploxererts complex, 0 to 30 percent slopes • Cropley, coastal-Xerorthents, landscaped-Urban land complex, 0 to 9 percent slopes • Elder fine sandy loam, coastal, 0 to 2

! 37! percent slopes • Fluvaquents-Riverwash complex, 0 to 5 percent slopes • Linne silty clay loam, 15 to 50 percent slopes • Linne silty clay loam, 9 to 15 percent slopes • Los Osos clay loam, 30 to 50 percent slopes • Pacheco silty clay loam, 0 to 2 percent slopes • Sapwi-Urban land complex, 0 to 50 percent slopes • Sapwi loam, 30 to 75 percent slopes LMSA (2) MSA (3) • Calcic Argixerolls, 30 to 75 percent • Balcom-Balcom, dark surface slopes association, 30 to 75 percent slopes • Cotharin-Rock outcrop-Tongva • Balcom silty clay loam, 30 to 50 complex, 30 to 75 percent slopes percent slopes • Cotharin-Talepop-Urban land • Gaviota stony sandy loam, 30 to 50 complex, 0 to 50 percent slopes percent slopes • Cotharin-Talepop association, 15 to • Talepop-Rock outcrop complex, 30 to 50 percent slopes 75 percent slopes • Cotharin-Talepop association, 30 to • Typic Haploxerepts, 15 to 30 percent 75 percent slopes slopes • Cotharin loam-Rock outcrop • Urban land-Xerorthents, landscaped complex, very bouldery, 30 to 75 complex, 0 to 5 percent slopes percent slopes • Urban land-Xerorthents, landscaped, • Cumulic Haploxerolls, 0 to 9 complex, rarely flooded, 0 to 5 percent slopes percent slopes • Danville-Urban land complex, 0 to 9 • Xerorthents-Urban land-Balcom percent slopes complex, 0 to 15 percent slopes • Danville-Urban land complex, 9 to • Xerorthents-Urban land-Balcom 15 percent slopes complex, 0 to 30 percent slopes • Gaviota-Rock outcrop association, • Zumaridge-Kawenga association, 30 50 to 100 percent slopes to 75 percent slopes • Kayiwish association, 0 to 9 percent • Zumaridge-Sapwi-Kawenga slopes association, bouldery, 30 to 75 percent • Kayiwish association, 2 to 30 slopes percent slopes • Zumaridge-Rock outcrop complex, • Linne-Los Osos-Haploxerepts bouldery, 30 to 75 percent slopes association, 30 to 75 percent slopes • Lockwood-Urban land complex, 0 to 15 percent slopes • Lockwood-Urban land complex, 0 to 9 percent slopes • Malibu-Chumash-Boades association, 15 to 50 percent slopes • Pachic Argixerolls, coastal, 30 to 75 percent slopes • Tongva-Cotharin-Rock outcrop complex, 30 to 75 percent slopes • Typic Haploxerepts, 30 to 50 percent slopes

! 38! • Urban land-Tongva complex, 0 to 15 percent slopes • Xerorthents, landscaped, 0 to 9 percent slopes HMSA (4) HSA (5) • Chumash-Boades-Malibu • Calcic Haploxerepts-Mollic association, 30 to 75 percent slopes Haploxeralfs association • Mipolomol-Topanga-Rock outcrop • Chumash-Boades-Malibu association, complex, 30 to 75 percent slopes 5 to 15 percent slopes • Mipolomol-Topanga association, 30 • Zumaridge-Rock outcrop-Sumiwawa to 75 percent slopes complex, very stony, 15 to 50 percent • Sumiwawa-Hipuk-Rock outcrop slopes complex, 30 to 75 percent slopes • Sumiwawa-Rock outcrop- Zumaridge complex, very stony, 30 to 75 percent slopes • Rock outcrop-Sumiwawa-Hipuk complex, 30 to 75 percent slopes • Topanga-Mipolomol-Sapwi association, 30 to 75 percent slopes • Zumaridge-Kawenga association, 15 to 50 percent slopes

4.3.3 Slope and Border Classification

Slope, measured in percent rise, was classified as HSA between zero and ten percent slope, HMSA between ten and twenty percent, MSA between twenty and thirty,

LMSA between thirty and forty, LSA between forty and fifty, and ZSA above fifty percent rise (Table 3). These are generally understood slope suitability levels for CHL due to their squat bodies, short legs, poor climbing ability, and need for flat basking and digging areas (Suarez, Richmond, and Case 2000, Moody et al 1991; Wiernasz and Cole

1995).

A human impact buffer was created inside the edge of the SMMNRA to indicate the direct interaction of humans with habitat boundary. Although roads and buildings are within the SMMNRA, for this research, only the outside boundary was considered. The buffer (500 meters) inside the edge of SMMNRA was treated as category 3 – MSA due in large part to the likely presence of disturbed vegetation. This assumed condition was

! 39! based on research of urban interface and suitability done with the Texas Horned Lizard

(Endriss et al 2007). The full impact of urbanization on CHL survivability and suitability is not well understood but Texas Horned Lizards can serve as a proxy for understanding that impact due to characteristics and needs similar to CHLs (Pianka and Parker 1975;

Brattstrom 1997; Endriss et al 2007).

This border buffer area represented the interaction zone between significant human impact outside of SMMNRA and less impact within the border. CHL do inhabit land outside of SMMNRA and are not limited to within its boundaries (CDFG/CIWTG

2007). Although conditions outside of the border do affect CHLs, this research was focused solely on what was within the SMMNRA boundaries. The conditions outside of

SMMNRA and what impact they may have on CHL habitat suitability was not part of this research.

Table 3 - Slope and Border Type Classification Table

Abbreviation Classification Value Description ZSA Zero Suitability Area 0 Slope - above 50 percent rise Border – N/A LSA Low Suitability Area 1 Slope – between 40 and 50 percent rise Border – N/A LMSA Low-Medium Suitability 2 Slope – between 30 and 40 Area percent rise Border – N/A MSA Medium Suitability Area 3 Slope – between 20 and 30 percent rise Border – within 500 meters HMSA High-Medium Suitability 4 Slope – between 10 and 20 Area percent rise Border – N/A HSA High Suitability Area 5 Slope – between 0 and 10 percent rise Border – N/A

! 40! 4.4 Reclassification

The reclassification stage consisted of two stages for each factor examined: pre- mapping the original data classification and reclassifying that map data to match the specifications in Section 4. For each factor (slope, soil, vegetation and border), these initial maps were used as the weighted overlay rasters. Some early results were apparent from examining this stage. These initial mapping results are described in detail below for each factor.

4.4.1 Vegetation

Initial vegetation data showed a core zone of chaparral and coastal sage scrub.

This CSS/chaparral core was broken up by pockets of disturbed vegetation (from human inhabitation or roads) and numerous riparian corridors. The largest areas of disturbed vegetation were located along the Malibu coastline, 101 freeway, cities within

SMMNRA, and a few cross-mountain roads (Topanga Canyon Boulevard, Malibu

Canyon Road, and Kanan Road). Due to SMMNRA’s complicated border, some areas of vegetation were almost entirely affected by disturbed vegetation, such as the easternmost projection. This geographic squeeze was also present between the Santa Monica

Mountains and Simi Hills. The greatest area of disturbance was along the coastline though, with Point Dume almost entirely a disturbed vegetation zone. Agricultural areas in Ventura County were visible as well as a few water features large enough to have shown on this scale. Upland trees occurred scattered throughout the spine of the mountains but were mostly found in the Topanga Canyon and Malibu Canyon areas.

Chaparral dominated much of the core area with a transition to CSS that became more

! 41! prevalent towards Ventura County.

These preliminary results agreed with the assumption in this research that most of

SMMNRA is at least minimally suitable for CHLs. The distribution of disturbed vegetation types followed very closely with roads and human communities. Even without road and town map overlay, human dominated areas could be clearly identified. Human impact is one of the core issues of conservation in SMMNRA. Disturbance of native vegetation often precedes the spread of invasive ants, as they prefer disturbed areas

(McIntyre 2003). Patch size of CHL habitat in these disturbed areas is likely to be small and fragmented due to human barriers like buildings and roads (Rose 1982). CHL have difficulty navigating between fragmented areas. Mating populations may become geographically isolated due to fragmentation Some of these areas may not meet the minimum population persistence area requirement of 100 acres (CDFG/CIWTG 2007).

These disturbed vegetation results agreed with the study’s designation of both border areas and disturbed vegetation as MPAs. Disturbed vegetation can still provide moderately acceptable habitat since some native vegetation is still present and harvester ants can inhabit these modified landscapes. Urban/built-up areas were found both intermixed with the disturbed areas as well as directly bordering them. This urban/disturbed interface is of great significance when planning potential expansions of the park area. These areas may be able to re-establish the presence of CSS/chaparral vegetation with human assistance (Bowler 2000). Finally, the area between Simi Hills and Malibu Canyon contained more prairie/meadow vegetation than any other area mapped. That area though also contained the largest single area of urban/built-up vegetation type, the City of Calabasas, which sits partially within the SMMNRA

! 42! boundaries. The Malibu flats area showed the same problem of significant human development. Agricultural land was found within Hidden Valley. These developed areas may pose a geographic reproductive barrier to CHLs that live near it (Martin and Murray

2010). Roads, buildings and other human structures have been found to restrict CHL movement and limit the range necessary to find unrelated mates.

Reclassified values showed a similar pattern to the original values, as would be expected given that the original symbology used a similar palette of colors. This was purposeful as it was easier to identify patterns using colors that were similar to nature rather than random colors. The reclassified values showed that there were only a few very small areas that were completely restricted in the weighted overlays. This map also showed that most of the SMMNRA range contained HSA vegetation (CSS/chaparral) with scattered fragmentation by lower suitability areas. Although these lower suitability areas coincided with human development areas, it is clear from this data that most of

SMMNRA remains the CSS/chaparral type that is the ideal vegetative habitat for CHLs.

! 43!

Figure 7 - Original Vegetation Map

! 44!

Figure 8 - Reclassified Vegetation Map

The rounded percent areas for reclassified vegetation values were .5% ZSA, 9%

LSA, .8% LMSA, 5% MSA, 15% HMSA, and 70% HSA. Note: Percents may not add to

100% due to rounding.

! 45! 4.4.2 Soil

The soil results were complicated due to the diversity of soil types present in

SMMNRA. The preliminary map was left in random unique values symbology to highlight the large number of types. These results did show that there was one large area of single soil type present in the eastern section of SMMNRA: Mipolomol-Topanga association, 30 to 75 percent slopes. This association was also found mixed with others to the west of Point Dume. A dominant series along the westernmost range into Ventura

County was the Chumash-Boades-Malibu association, 30 to 75 percent slopes. The

Malibu coastal Malibu flats area contained areas of Calcic Argixerolls, 30 to 75 percent slopes; Cropley, coastal-Urban land-Haploxererts complex, 0 to 30 percent slopes; and

Urban land-Xerorthents, landscaped, complex, rarely flooded, 0 to 5 percent slopes.

These are low suitability areas for CHL due to urban influence and textural factors. This low suitability soil area was also a low-to-medium suitability vegetation area due to human impact. This coordination of poor soil and mixed poor vegetation was not surprising given the influence vegetation disruption can have on soil type. Found just inland though was the Chumash-Boades-Malibu association, 30 to 75 percent slopes, that was most common along the coast-face of the mountains. Although the coastal flats are considered poor soil habitat for CHLs, the soil just inland along the mountains was mostly high-medium and high suitability.

Soil suitability becomes a lesser factor if the soil patch is small and able to be traversed by a CHL moving from one range to another (Rose 1982). Interspersed areas of low suitability are acceptable as long as the percent of acceptable soil type within an area is greater than 75% (CDFG/CIWTG 2007). This allows for migration through low

! 46! suitability areas. Although home range migration and long-distance travel is rare, finding new territory is important for CHLs experiencing push factors such as human impact or development, invasive ant spread or wildfires. Even the LSA soils present are able to sustain CHLs in these situations until more suitable habitat can be found.

Another area of note was the Australia-shaped patch in the northwestern region that consisted of Tongva-Cotharin-Rock outcrop complex, 30 to 75 percent slopes;

Cotharin clay loam, 30 to 75 percent slopes; Cotharin loam-Rock outcrop complex, very bouldery, 30 to 75 percent slopes; Cotharin-Talepop association, 15 to 50 percent slopes; and Cotharin-Talepop association, 15 to 50 percent slopes. It was enclosed on three sides by the more extensive Mipolomol-Topanga association, 30 to 75 percent slopes and

Chumash-Boades-Malibu association, 30 to 75 percent slopes. This “Cotharin cluster” corresponded to an area along the Ventura/Los Angeles county line roughly from Circle

X Ranch to Rocky Oaks, encompassing Hidden Valley, Malibu Springs and Little

Sycamore Canyon Road. These Cotharin soil types were mostly LMSAs due to high levels of loam without a significant gravel/sand percent. These areas are habitable though because they also contained chaparral/CSS vegetation and acceptable slope profiles. As with the Malibu flats, the “Cotharin cluster” is able to support CHL lifecycle needs but is not as suitable as other soil types. Suitability is not restrictive unless it is a ZSA. Features like water and wetlands are unable to support CHLs and as such cause that area to be automatically considered ZSA. This variability of factors (i.e. vegetation is HSA but soil is LMSA) was found throughout the region. The implications of this could have serious impact on planning for CHL conservation. It is unlikely to find many areas that are fully

HSA in all categories so any conservation or research plans would need to account for a

! 47! wider range of potential acceptable areas. This data is only an initial survey of known suitability characteristics, not an extensive catalogue of high suitability areas.

In the reclassified map, all soil types were assigned a value from zero to five, according to the classification scheme. The simplification from the original soil types to 6 habitat suitability categories made identifying patterns and suitability areas significantly easier. The “Cotharin cluster” was easily visible as a large LSA/LMSA zone that also connected with the northern protrusion of SMMNRA that includes Cheseboro and Palo

Comado Canyons. This connecting corridor between the “Cotharin cluster” and Simi

Hills contained the Linne-Los Osos-Haploxerepts association, 30 to 75 percent slopes and

Cotharin-Talepop association, 15 to 50 percent slopes. The connector was broken up by a distinct section of HSA soil (Calcic Haploxerepts-Mollic Haploxeralfs association, 30 to

75 percent slopes) found within the protected area of northern Malibu Creek State Park.

On the non-classified map, this distinction between HSA and LMSA was not as visible.

The reclassification highlighted the greatly variable distribution of soil types.

Due to few soil series being classified as HSA, the soil results were comparably less favorable than the vegetation results. The choices behind the soil series classification scheme had a serious impact on the results. As with all habitat suitability modeling, the classification scheme of values affects the results and subsequent analysis. For example, reclassification of Chumash-Boades-Malibu association, 30 to 75 percent slopes;

Mipolomol-Topanga-Rock outcrop complex, 30 to 75 percent slopes; and Mipolomol-

Topanga association, 30 to 75 percent slope as HSA rather than HMSA would increase the total HSA amount. These were classed as HMSA though due to their composition and provided slopes. More research is needed on the classification of suitable soils for CHLs

! 48! before more comprehensive conservation plans could be made.

Figure 9 - Original Soil Series Map

! 49!

Figure 10 - Reclassified Soil Series Map

! 50!

The rounded percent areas for reclassified soil values were 1% ZSA, 15% LSA,

28% LMSA, 9% MSA, 45% HMSA, and 2% HSA. Note: Percents may not add to 100% due to rounding.

4.4.3 Slope

Slope results on the initial map were already partially classified by manual breaks.

This slope represents percent rise, rather than degrees. For CHLs, percent rise is critical for their mobility because they cannot easily climb slope percent rises greater than 50.

This is due to their short limbs and small size. Even a small percent rise represents a significant increase in difficulty for their movement (Pianka and Parker 1975).

Reclassification was a matter of assigning the zero-to-five values to each existing break since slope values were grouped initially. Slope percent results extended throughout the study area and did not contain any “no data” values. CHLs can inhabit a wide range of slope percents although lower slope percents are more suitable. CHL are often found in low suitability slope percents due to evasive action from predators, groundcover selection for thermoregulation and the search for mates (Pianka and Parker

1975). For the purposes of this study, low slope percents were deemed high suitability but future research is needed regarding small-scale slope variations that are not available in

DEM data and on what scale slope should be examined for CHL conservation (Mayer and Cameron 2003).

High suitability slope areas were found along the Malibu coastal flats, Point

Mugu, Hidden Valley, parts of northern and southern Malibu Creek State Park and Tapia

Park. Higher slope percents (low suitability) were found along the spine of the Santa

! 51! Monica Mountains and along its many canyons. Some of these places were prohibitively steep for CHL, such as parts of the southern Malibu Canyon/Piuma Road area. There were varying slopes present but the significant presence of ZSA slopes may restrict CHL movement between otherwise acceptable areas (Rose 1982). CHL cannot climb rock faces and rely on short bursts of running on flat to moderate slopes to escape predators

(Pianka and Parker 1975). High slope percents preclude that essential survival strategy.

For this study, slope percents were separated into simple 10% categories although a more complex and accurate picture may be derived from making the breaks non-equal and setting slopes from 0% to 30% as HSA. These equal breaks results showed that, in terms of simple slope classification, the highest suitability areas did not always match with highly suitable characteristics in other categories (i.e. Malibu flats slope – HSA, Malibu flats soil and vegetation – ZSA to LMSA).

The presence of higher suitability areas (3-5) throughout SMMNRA, especially in the northern areas, indicated an increased likelihood of larger patch sizes for mate selection. Mate selection occurs in larger patches than normal activity ranges, leading to expansion and exchange of territories (Rose 1982; Goldberg 1983). Slope does act a restricting factor for range movement but from this data, the northern corridor allowed for a relatively flat expanse for mating movement. This area was also closest to the human impact areas along the Conejo and San Fernando Valleys. Human impact may make these areas less suitable, even though the slope was within acceptable bounds, due to soil or vegetation disturbance.

! 52!

Figure 11 - Original Slope Percent Map

! 53!

Figure 12 - Reclassified Slope Percent Map

! 54! The rounded percent areas for reclassified slope values were 0% ZSA, 2% LSA,

16% LMSA, 35% MSA, 30% HMSA, and 17% HSA. Note: Percents may not add to

100% due to rounding.

4.4.4 Borders

Border results on the initial map were already classified into one value. Border results extended inside the study area and were not unexpected given the erratic border of

SMMNRA. Initial data showed that the easternmost protrusion of SMMNRA (from 405 freeway to 101 freeway heading east) was entirely within the border zone. This whole extension was considered MSA. The 500-meter border zone within SMMNRA represents a zone of human impact where, even with other high suitability characteristics, the survivability of CHLs could be greatly reduced due to predation by domesticated cats, road accidents, invasive ant preference for human areas and habitat destruction (Holway

2005; McIntyre 2003). Internal borders, trails and roads were not examined in this research, although roads and internal development borders within SMMNRA should be affected by the same survivability concerns as the exterior border.

! 55!

Figure 13 - Original SMMNRA Buffer Map

Figure 14 - Reclassified SMMNRA Buffer Map

! 56!

4.5 Landscape Metrics

After the classification of variables and the formation of the weighted overlay models, selected class-level landscape metrics were analyzed to assess habitat fragmentation and relationships between different types of suitability areas. Landscape metrics must be used carefully in order to ensure appropriate interpretations. There are limitations of each measure, potential ranges of data values, and shifts in the ranges of those values that should be accounted for in each type of metric. These differences in metrics can represent different structural characteristics of a landscape (Hargis et al 1998;

McGarigal et al 2012). Landscape metrics were used to compare quantitative fragmentation levels within and between each suitability area type, as well as distribution patterns of different suitability types. Composition and configuration of suitability areas are both important to address as indicators of fragmentation (Swenson and Franklin 2000;

McGarigal et al 2012).

Landscape metrics cannot solely describe the suitability or fragmentation of an area but in conjunction with the models and other known habitat data, some interpretation of suitability and fragmentation patterns could be made (Swenson and Franklin 2000). If habitat fragmentation in the SMMNRA is occurring, it may be due to increasing the number of landscape pieces, decreasing interior habitat area, increasing the extent of landscape edges or increasing isolation of residual patches (Li et al. 1993). All of these factors involve changes in the composition and/or configuration of the landscape.

Class level metrics were used in this analysis because they represented the amount and spatial distribution of patch types and could be interpreted as fragmentation indices

! 57! (McGarigal et al 2012). FRAGSTATS was used to calculate all of the metrics used in this research (McGarigal et al 2012). Four class level metrics were examined for each habitat suitability category on all three models. The metrics calculated were percentage of landscape area (PLAND) occupied by each habitat suitability category, area-weighted mean Euclidean nearest neighbor distance (ENN_AM) between habitat suitability category patches (which includes area-weighted mean (AREA_AM)), percent of like adjacency (PLADJ) among pixels in the landscape, and interspersion/juxtaposition index

(IJI). These four metrics are described in further detail below.

1. The percentage of class types on the landscape (PLAND) was measured as the

primary indicator of type composition. PLAND gave the percentage of each

cover type on the landscape out of 100 percent. This metric made it easy to

compare how much of a suitability type was present in the SMMNRA on each

model. The equation for PLAND index was:

where Pi = proportion of the landscape occupied by patch type (class) i. aij =

area (m2) of patch ij. and A = total landscape area (m2) (McGarigal et al

2012). This equation summed the areas of all patches of the same type and

divided it by the total landscape area, then multiplied it by 100 to get a

percentage (McGarigal et al 2012). The possible range values were between 0

and 100. If the resulting value was close to 0 it meant that class type was rare

on the landscape, and if the value approached 100 it meant one class type

! 58! covered the whole landscape (McGarigal et al 2012).

PLAND served as a fundamental measures of landscape composition,

indicating how much of each suitability area type existed within the

SMMNRA. A key effect of habitat fragmentation is habitat loss so it was

important to know how much of each habitat suitability type existed. This was

the starting point for analyzing patterns in the models and interpreting

pressures in different areas. CHL do not need a large patch of one single type,

but can intermix and transition between HSA, HMSA, and MSA with relative

ease. Permanent habitat suitability is for HSA characteristics although both

high HSA and HMSA PLAND percents are beneficial for CHL survival

(Pianka and Parker 1975, Brattstrom 1997; CDFG/CIWTG 2007).

2. The Euclidean nearest neighbor area-weighted mean distance (ENN_AM)

measured the distance to the nearest neighbor, aij, for the patch type ij in

meters (ENN), and compensated for differences in the distribution of patch

sizes, which can affect inter-patch distances in a spatially limited landscape.

The area-weighted mean (AREA_AM) portion of this metric used this

equation:

The ENN for each class was divided by the area of all nearest neighbor

distances of the same class on the landscape, and then multiplied by the

! 59! number of neighbors (McGarigal et al 2012). Units were in meters, the native

unit of the rasters used.

The ENN_AM was used to describe if nearest neighbors (of same type)

were close together or far apart on a landscape. This value indicated the

distance between certain suitability type patches and other patches of the same

type, but the distance to large similar type patches were given more weight. In

terms of fragmentation, shorter distances between HSAs and HMSAs were

more suitable as they represented more intact patches of suitable habitat of

appropriate size (CDFG/CIWTG 2007).

3. The percentage of like adjacencies on the landscape (PLADJ) was used as

an indicator of contiguity with similar connected suitability areas. This

measure uses pixels rather than whole patches to assess adjacency. The

equation for PLADJ index was:

where gii = number of like adjacencies between pixels of patch type (class) i

based on the double-count method and gik = number of adjacencies between

pixels of patch types (classes) i and k based on the double-count method

(McGarigal et al 2012). In the double-count method, each pixel adjacency is

counted twice and the order of pixels is preserved with two exceptions for

background and boundary values if present. Only SMMNRA boundary values

occurred as an exception in this analysis. Percentage of like adjacencies ! 60! (PLADJ) was computed as the sum of the diagonal elements (i.e., like

adjacencies) of the adjacency matrix divided by the total number of

adjacencies. A landscape containing larger patches with compact shapes will

contain a higher proportion of like adjacencies than a landscape containing

smaller patches and more complex shapes. This index provided an effective

measure of class-specific aggregation that isolated the dispersion component

of aggregation (McGarigal et al 2012).

PLADJ values indicate whether a specific class cell (suitability type) is

likely to have same value adjacent cells. Although the same suitability type is

not needed for CHL habitat suitability (such as HSA and HMSA intervening

areas), this index provided a measure of fragmentation of each suitability type

(Pianka and Parker 1975, Brattstrom 1997; CDFG/CIWTG 2007). It showed if

HSAs were more or less likely to be bordered by other HSAs as compared to

other types. For this analysis, high HSA PLADJ values were beneficial to

CHL habitat suitability but only when considered in conjunction with other

metrics.

4. The interspersion and juxtaposition index (IJI) was used as a spatial

configuration metric that isolated the interspersion or intermixing of patch

types. IJI provided information on how intermixed different patch types were

with other cover types. The formula for IJI was:

! 61!

In this formula, m equaled the number of class types on the landscape, and

eik was the total length in meters of the edge between class i and k (McGarigal

et al 2012). The observed interspersion was divided over the maximum

possible interspersion for the given number of patch types (McGarigal et al

2012). Range was 0 < IJI < 100 and in percent units. IJI approached 0 when

the corresponding patch type was adjacent to only 1 other patch type and the

number of patch types increases. IJI was 100 if the corresponding patch type

was equally adjacent to all other patch types (i.e. maximally interspersed and

juxtaposed to other patch types) (McGarigal et al 2012).

IJI was used to measure the spatial intermixing of different suitability

areas. Interspersion only accounts for how often each suitability area is

adjacent to each other suitability area and not by the size, contiguity or

dispersion of patches (McGarigal et al 2012). This metric was used with

PLADJ and ENN_AM to assess suitability area continuity and distribution of

suitable habitat areas. In terms of fragmentation, high levels of interspersion

may indicate fragmented habitat, although care must be taken in interpreting

these results, as the interspersed area may also be suitable habitat. For

example, some HSA IJI values included interspersion with HMSAs that were

part of normal and acceptable habitat variation in CHL (CDFG/CIWTG

2007). On the whole though, IJI indicated the rate of intervening habitat

! 62! suitability areas within each suitability category. IJI analysis required

assessing visual patterns on the models and discerning what suitability area

types were actually interspersed. IJI must be interpreted in respect to other

metrics and patterns to present an accurate view of fragmentation.

! 63! 5.0 - RESULTS

5.1 Expectations

This research set out to use CHL suitability factors detailed in the literature related to CHL habitat selection to create suitability maps. The results presented here are not meant to be a comprehensive analysis of all habitat factors for CHL or a predictive model for the location of actual CHL populations. The results provide a starting point for further research and planning of CHL species management.

The expected result of this research was that highly suitable areas (HSAs) would appear fragmented due to differences in slope percent, vegetation and soil type across

SMMNRA. Slope should have affected the distribution of HSAs, as CHL do not generally have long-term inhabitation of severe slope percent areas (greater than 40%) due to their need for open, flat basking areas (Suarez, Richmond, and Case 2000, Moody et al 1991; Wiernasz and Cole 1995). Vegetation type should have had less of an impact on HSA distribution as much of SMMNRA is within the acceptable CHL vegetation profile, although HSAs should contain coastal sage scrub and/or chaparral. Human impact areas, considered the border of the park for the purposes of this study, were considered less than ideal habitat for two reasons: modifications of the habitat due to human interaction and invasive ants preference for areas of disturbance (McIntyre 2003;

Suarez, Richmond, and Case 2000, Moody et al 1991; Wiernasz and Cole 1995). Overall, it was expected that HSAs occur in fragmented locations due to variations in soil type, slope and vegetation type.

! 64! 5.2 Suitability Models

All major HSAs based on each of the four individual environmental factors were given approximate boundaries and general names to clarify their location on the map. The extent of these HSAs varied depending on the model. These reference points can be found on the official NPS map of SMMNRA.

Table 4 – Approximate Landmark Boundaries of the Four Identified HSAs

Northern Malibu Creek State Park HSA 2 kilometers south of 101 freeway (north), Goat

Buttes (west), Tapia Park (south) and King

Gillette Ranch (east).

Southern Malibu Creek State Park HSA Backbone Trail (north), Castro Crest (west), 2

kilometers north of Highway 1 (south), and

Malibu Creek (east).

Rambla Pacifico HSA Backbone Trail (north), Piuma Road (west),

Rambla Pacifico (south), and the intersection of

Saddle Peak Road and Tuna Canyon Road (east).

Point Mugu State Park HSA Wood Canyon (north), Mugu Peak (west),

Serrano Valley (south), and Boney Point (east).

HSAs were not completely composed of HSA designated land but were generally greater than 75% HSA (CDFG/CIWTG 2007). For the purposes of simplifying the description of fully functional and suitable areas for CHLs, extensive areas of contiguous

(non-fragmented) combinations of HSAs and HMSAs (less than 75% but greater than

25% HSA) are herein referred to as HSA/HMSA complexes.

! 65!

Figure 15 - Map of the Four Major Identified HSAs in SMMNRA

! 66! 5.2.1 Suitability Model 1

Model 1 included the soil, vegetation, border and slope layer rasters. The weighting was vegetation 30%, soil 30%, slope 25% and border 15%. Any areas that did not contain data were given “restricted” values and appear on the model automatically as

ZSA. Only 4.29% of the areas studied were ZSA and .02% (rounded to 0%) was LSA.

The main reason for a ZSA label was “no data,’ while the second most common reason was the presence of water features. LMSAs were only 3.36% of the study area. These occurred mostly along the lagoon of Point Mugu, a section of Topanga Canyon and patches around Calabasas.

The Point Mugu lagoon area value was low due to poor soil, vegetation and border values. The slope in this area was a HSA but the other low scoring factors drove down its final value. CHL do not inhabit beach/lagoon areas unless forced by fires into those areas. The Topanga Canyon LMSA patch (at the city of Topanga) was due to disturbed vegetation and soil. The slope and border ratings were high in that patch but, due to the weighting, the undesirable soil and vegetation values caused the final value to be LMSA. The Calabasas patches were deemed LMSA due to disturbed soil and vegetation, as well as border proximity. The slope was highly suitable in those patches.

These LMSA patches show the impact of human development within SMMNRA.

Calabasas and Topanga are cities within SMMNRA that are experiencing expanding development. Malibu also showed pockets of LMSAs but that was mainly because of coastal soil types and improper vegetation, although it did include a moderate level of disturbance. Much of the Malibu flats area around Point Dume was considered MSA for slope and border but this area had poor soils and urban/built-up vegetation types.

! 67! The disturbance of soil and vegetation had a severe impact on CHL habitat suitability. Areas with disturbed or urban characteristics scored very low on soil and vegetation. This caused many areas to be MSA through a mix of very high and very low scores. This impact could be examined by changing the weighting, although for this study only three weighting equations were chosen and balanced fairly equally with border the least important (Hernandez et al 2006).

HSAs included both northern and southern Malibu Creek State Park and Rambla

Pacifico. These were the largest identifiable HSAs on Suitability Model 1. HMSAs

(66.27%) were prevalent and spread throughout the horizontal axis of SMMNRA. The largest of the nearly intact HMSAs were Topanga State Park, Zuma/Trancas Canyon and

Point Mugu State Park. These three are all actively protected areas as either NPS land or private parkland. Through this model, protected areas were easily identifiable; not because they were designated as such on these layers but because that protection has ensured the continued presence of proper soils and native vegetation needed by CHLs.

This data could be used in conservation planning to help prove the value of these undeveloped, protected areas for native wildlife (Cassidy and Gore 2000; Holway 2005).

High levels of fragmentation in chaparral/CSS vegetation may cause CHLs to be restricted to small, untenable habitats (Brooks et al 2002). Northern Malibu Creek State

Park HSA, which directly bordered the city of Calabasas, showed distinct boundaries between excellent conditions within the Park and drastically lower scores to the north due to human disturbance. Variations in slope caused some fragmenting of the HSA/HMSA complex there but the topography included canyon passages that allow for CHL movement between valley areas.

! 68! Overall, this weighting did allow for distinct and identifiable HSA regions, which was the purpose of this study. According to these results, those HSAs were likely to have significant CHL populations or at least highly suitable habitat for them.

Table 5 – Suitability Model 1 Total Area Type Percent

TYPE PLAND (%)

ZSA 4.2936

LSA 0.0161

LMSA 3.3591 MSA 23.1082

HMSA 66.269

HSA 2.9539

! 69!

Figure 16 – Suitability Model 1

Weighting: Vegetation 30%, Soil 30%, Slope 25% and Border 15%

! 70! 5.2.2 Suitability Model 2

Model 2 included the soil, vegetation, border and slope layer rasters. The weighting was vegetation 40%, soil 30%, slope 20% and border 10%. Any areas that did not contain data or were not digitized were given “restricted” values and appeared on the model automatically as ZSA. Only 4.29% of the areas studied were ZSA and .3% was

LSA. The main reason for a ZSA label was “no data,’ while the second most common reason was the presence of water features. LMSAs were 5.91% of the study area. This is higher than in Suitability Model 1 (3.35%). These occurred in the same areas as

Suitability Model 1 (lagoon of Point Mugu, Topanga and Calabasas) but included in this model a large area in Hidden Valley.

The Point Mugu lagoon area value was still low due to poor soil, vegetation and border. The Topanga Canyon LMSA patch (at the city of Topanga) from Model 1 remained LMSA for the same reasons but this time was bordered by more HSAs from the increased vegetation weighting. Previously in Model 1, the Topanga LMSA was bounded by mostly HMSA. That showed a distinction between human impact and undisturbed land. On this model, that distinction was more noticeable as the Topanga area was mostly

HSA/HMSA except for the town center. The Calabasas patches were again deemed

LMSA due to disturbed soil and vegetation, as well as border proximity. The slope was still highly suitable in those patches but was weighted less in this model. Areas of human development were again identifiable on the model, mostly as LMSAs, because of their disturbed soil and vegetation profiles. Malibu, Topanga and Calabasas were the centers of human habitation within SMMNRA. The continued presence of human disturbance may change the surrounding area’s profile even without the building of more houses or roads.

! 71! Predation by felines, human-caused fires, increased invasive ant spread and fragmenting of otherwise high suitability habitat may cause these LMSAs to expand over time

(McIntyre 2003). Fewer areas were MSAs this time due to the heavier weighting of vegetation. Along with the decrease in HMSAs came a significant rise in HSAs, through reclassification based more on vegetation. Most of the SMMNRA was HMSA or HSA for vegetation and thus, when weighted more heavily, there was a corresponding rise in

HMSA/HSA model values. As this was only one of many potential weighting equations, future research could examine if vegetation should be counted as high as 40%, like in this model. This was only a baseline examination of habitat suitability as a guide to future research questions.

HSAs still included northern and southern Malibu Creek State Park as well as

Rambla Pacifico but were joined by a scattering of HSA fragments throughout the Park, which included a new large HSA in Point Mugu State Park. HMSAs were lower this time at 59.1%, attributed to the increase in HSAs. The nearly intact HMSAs of Model 1

(Topanga State Park, Zuma/Trancas Canyon and Point Mugu State Park) now included a fairly random distribution of HSAs within them. Appearing far clearer in this model was the “Cotharin cluster” that was identified on the soil reclassification map. The “Cotharin cluster” was a LSA/LMSA soil series. Even though this area contained mostly HSA vegetation, there were scattered areas of urban/disturbed vegetation as well as agriculture.

Due to the heavier vegetation weighting and poor soil profile, this “Cotharin cluster” was entirely devoid of HSAs. Variations in slope again caused some fragmenting of the

HSA/HMSA habitats, but it followed the same pattern as Model 1. The decrease in slope weighting though also caused an increase in HSAs upgraded from HMSAs, contributing

! 72! to the higher percent of HSAs.

Overall, this weighting showed more HSA regions scattered throughout

SMMNRA in addition to the ones from Model 1. This was a significant increase in total

HSAs. According to these results, more areas were HSA/HMSAs when vegetation is weighted at 40% instead of 30%.

Table 6 – Suitability Model 2 Total Area Type Percent

TYPE PLAND (%)

ZSA 4.2936

LSA 0.3212

LMSA 5.9126

MSA 18.616

HMSA 59.0974

HSA 11.7591

! 73!

Figure 17 – Suitability Model 2

Weighting: Vegetation 40%, Soil 30%, Slope 20% and Border 10%

! 74! 5.2.3 Suitability Model 3

Model 3 also included the soil, vegetation, border and slope layer rasters. The weighting was vegetation 30%, soil 30%, slope 30% and border 10%. There were areas in the soil layer that did not contain data and were given “restricted” values. These automatically appeared on the model as ZSA. Areas with no data were excluded as the scope of this study only covers known factors. The main reason for a ZSA label was “no data,” while the second most common reason was again the presence of water features.

All three models showed this pattern. Only 4.16% of the areas studied were ZSA and

.04% was LSA. LMSAs were 3.43% of the study area. This was similar to Model 1

(3.35%) but lower than Model 2 (5.91%). These occurred in the same areas as Model 1

(lagoon of Point Mugu, Topanga and Calabasas) but unlike Model 2, Hidden Valley was not considered LMSA in this model. Hidden Valley appeared as MSA on this model.

When vegetation was more heavily weighted (40% in Model 2), Hidden Valley changed from MSA to LMSA that indicated less suitable vegetation (agriculture) in that area.

Hidden Valley contained the largest section of agricultural land within

SMMNRA. As agricultural vegetation was considered LSA, higher weighting of vegetation caused this difference. CHLs can live in agricultural land but due to agricultural machines, soil disturbance and pesticide usage; this land type is very dangerous to CHLs (Hellgren et al. 2010). Agricultural pesticides kill harvester ants and can directly poison CHLs (Pianka and Parker 1975). It was not known, for the purposes of this study, whether pesticides are in fact used in Hidden Valley but the risk of blunt trauma by agricultural machinery remains. Point Mugu lagoon and Topanga Canyon (at the city of Topanga) LMSAs that were present in both Models 1 and 2 remained such in

! 75! this model. Model 1 and Model 3 appeared very similar since both models had a 30% vegetation weight. All three models showed a distinction between areas of human development and non-developed areas through soil disturbance and urban/built-up vegetation profiles. Methodological questions remain as to whether vegetation and soil should be weighted higher than 30% given the critical role that urban/disturbed and built- up landscapes play in determining low suitability areas (Hernandez et al 2006). These areas do not preclude CHL inhabitation but they have increased risk of CHL mortality due to more frequent human contact, in addition to poor suitability profiles (Hellgren et al. 2010; Brooks 2002). The occurrence of LSA/LMSA habitats within SMMNRA according to the data appeared to be related in almost every instance to the presence of human development. A notable exception occurred in Point Mugu where lagoon/wetland conditions cause the low suitability.

The three identified HSA areas were northern and southern Malibu Creek State

Park and Rambla Pacifico but they were smaller and intermixed with more HMSA compared to Model 1. HSAs were less dense on this overlay, which led to smaller, more fragmented HSAs. This reflected the role that increased slope percent weighting played in the calculations. These areas indicated a decrease of HSA conditions when slope was given greater weighting. HMSAs were 66.09% of the area, which reflected similar results found in Model 1. The percentage results in Model 1 and Model 3 were within 1% of each other. Model 1 and Model 3 provided nearly the same percent results, but Model 1 showed less fragmentation though as evidenced by a higher PLADJ value and lower IJI value. The decrease in visual HSA clumping represented a shift between HSA and

HMSA values. A slightly higher number of HSAs were demoted to HMSAs as there were

! 76! HMSAs upgraded to HSAs on Model 3. The distribution (configuration) of HSAs was more scattered as well. The difference between them was a 5% increase of slope weight in Model 3 and on Model 1 that 5% increase was placed on border instead. On both models, vegetation and soil were held at 30%. The fragmenting of HSAs that were intact in Models 1 and 2 was unexpected. Considering vegetation and soil at 30% each, slope percent was not expected to have such an impact at a 5% increase in weighting. Slope percent may play a larger role in habitat suitability but these models are only three possibilities of the weighting structure. Field surveys involving slope suitability may be needed to the further clarify these findings.

Overall, this weighting showed HSA regions more fragmented than in Model 1.

The percent of HSAs remained nearly the same as Model 1 though due to scattering of

HSA patches.

Table 7 – Suitability Model 3 Total Area Type Percent

TYPE PLAND (%)

ZSA 4.1575 LSA 0.0403

LMSA 3.425

MSA 23.6251

HMSA 66.0999

HSA 2.6523

! 77!

Figure 18 – Suitability Model 3

Weighting: Vegetation 30%, Soil 30%, Slope 30% and Border 10%

! 78! 5.3 Metrics

The metrics described in section 4.4 (PLAND, ENN_AM, PLADJ, and IJI) were calculated for each model. Each model’s results are presented separately here and then compared in the discussion section. Area-weighted mean (AREA_AM) was included in each model’s table as it was part of the Euclidean nearest neighbor (ENN_AM) calculation and for reference. A comprehensive table including all three models and all class level metrics is located in Appendix C.

5.3.1 Suitability Model 1 Metrics

Model 1 showed a low percent of HSAs located within large HMSAs (Table 8).

HMSA had the highest IJI (45.82%) due to mixing with MSA. Along with visual interpretation of the model, these metrics suggested that HSAs are mostly surrounded by either other HSA patches or HMSAs that are also very suitable habitat. The ENN_AM for Model 1 HSAs was 266.7 meters. Along with the low PLAND and high PLADJ, this suggested that the few HSA patches formed compact individual patches. This could indicate clumping of highly suitable habitats due to the lower slope weighting. Model 1 did show three identifiable HSAs but they were small and a low percent of total land.

Fragmentation of each HSA was not severe, mostly due to HMSA interspersion, but the

HSAs were highly isolated on visual assessment. Although isolated, the HSAs were surrounded by HMSAs. If assessed with HMSAs, the HSA/HMSA complex PLAND was

69.22% of the SMMNRA, an expected and positive result for CHL habitat suitability.

! 79!

Table 8 – Model 1 Metrics

TYPE PLAND AREA_AM ENN_AM PLADJ IJI

percent sq. meters meters percent percent

ZSA 4.2936 554.3549 257.3207 71.109 63.7575

LSA 0.0161 1.2 3151.0068 0 41.3718

LMSA 3.3591 45.4728 298.1434 42.5757 48.086

MSA 23.1082 676.7403 208.9997 55.7139 33.5137

HMSA 66.269 35268.5046 200.756 82.8303 45.8265

HSA 2.9539 53.3486 266.6996 29.9727 12.1779

5.3.2 Suitability Model 2 Metrics

Model 2 showed a higher percent of more intact HSAs located mostly within large HMSAs (Table 9). Interspersion and juxtaposition was higher for HSA in this model than in Models 1 and 3. HMSA IJI (60.62%) in model 2 was mostly due to MSA interspersion but also with increased HSA adjacency as shown on the model. The

ENN_AM for Model 2 HSAs was 210.17 meters. Along with visual interpretation of the model, these metrics suggest that HSAs covered more total area and were more likely to have similar adjacent patches but were also spread more evenly throughout the

SMMNRA. There was an increase in HSA IJI at 16.23% although upon visual inspection, most of this was due to HMSA interspersion. Model 2 showed three large HSAs but also extensive areas of intermixed HSAs and HMSAs. As a result, the extent of the largest three HSAs was more difficult to delineate than the other models. The higher HSA

! 80! PLADJ at 40.21% suggested more neighboring patches which may support CHL mating and movement needs (Brattstrom 1997). If assessed with HMSAs, the HSA/HMSA complex PLAND was 70.86% of the SMMNRA, a slight increase over Model 1.

Although the total HSA/HMSA PLAND did not increase much, that area was more evenly mixed with HSAs. This wider distribution caused by higher density and percent of

HSAs indicated more highly suitable and mostly intact habitat for CHLs.

Table 9 – Model 2 Metrics

TYPE PLAND AREA_AM ENN_AM PLADJ IJI

percent sq. meters meters percent percent

ZSA 4.2936 554.3549 257.3207 71.109 73.1539

LSA 0.3212 36.3668 519.687 52.2613 54.8118

LMSA 5.9126 117.2943 251.3657 50.6416 59.775

MSA 18.616 270.507 212.0794 54.6649 49.2525

HMSA 59.0974 29293.1335 201.4311 76.0188 60.6245

HSA 11.7591 71.5019 210.1708 40.2128 16.2339

5.3.3 Suitability Model 3 Metrics

Model 3 showed a low percent of HSAs located within large HMSAs (Table 10).

This was very similar to Model 1 metrics in composition but not in configuration. HMSA

IJI (46.28%) was again mostly due to MSA interspersion. The ENN_AM for Model 3

HSAs was 288.15 meters. Along with visual interpretation of the model, these metrics suggest that HSAs were more fragmented and the HSA patches were farther from each other. This resulted in a significant decrease in HSA PLADJ compared to Models 1 and

! 81! 2. The Model 3 HSA patches were surrounded by more non-HSA patches than in both

Models 1 and 2. Along with the low PLAND, this suggests that the few HSA patches were smaller and scattered (IJI 13.25%) so increased distance and lower size caused the

ENN_AM value to increase. Model 3 did show three identifiable HSAs but unlike Model

1 they were more difficult to delineate. The configuration difference between Models 1 and 3 was best shown here since the HSA PLANDs were similar but visually Model 1 showed much denser HSAs. This model’s HSA PLADJ was lower than Model 1 as a result of increased slope weight fragmentation. The fragmentation of suitable land was minimal, again mostly due to HMSA interspersion. The HSA/HMSA complex PLAND was lower than Models 1 and 2. It totaled 68.75% of the SMMNRA.

Table 10 – Model 3 Metrics

TYPE PLAND AREA_AM ENN_AM PLADJ IJI

percent sq. meters meters percent percent

ZSA 4.1575 585.4951 241.6921 72.662 64.8065

LSA 0.0403 3.08 1551.0527 18 45.6402

LMSA 3.425 43.5831 273.8319 40.0612 49.2433

MSA 23.6251 691.8823 207.783 55.5859 33.4471

HMSA 66.0999 34667.6968 200.5404 82.3294 46.2756

HSA 2.6523 7.8358 288.1471 17.0925 13.2473

! 82! 6.0 - DISCUSSION

Both visual model analysis and metrics showed evidence of fragmentation in highly suitable habitats on all three models. This was expected considering the literature available on habitat fragmentation patterns, the SMMNRA, and CHL habitat requirements (Brattstrom 1997; Beauchamp et al 1998; Swenson and Franklin 2000;

Burrow et al 2001; Fisher, Suarez and Case 2002; Sherbrooke 2003; Endriss et al 2007;

CDFG/CIWTG 2007; Leaché et al 2009). Interpretation of the suitability and fragmentation results relied on the differences in weighting as well as the classification scheme used for each factor. The implications of classification and weighting are discussed in section 6.1. The three research questions posed in section 1.2 are addressed separately in sections 6.2 through 6.4.

6.1 Patterns and Metrics

Vegetation and soil type were the most influential in determining habitat suitability, as they were more heavily weighted than slope and border. In all three models, vegetation and soil were given the highest percent weightings. Although only four factors (soil, vegetation, slope and border) were assessed, many other factors affect habitat suitability (Sherbrooke 2003). Of the four selected in this research though, vegetation and soil type were deemed the most influential based on the literature

(Brattstrom 1997; Beauchamp et al 1998; Burrow et al 2001; Fisher, Suarez and Case

2002; CDFG/CIWTG 2007; Leaché et al 2009). In particular, the presence of aquatic soil or vegetation types was an automatic ZSA designation. This automatic designation affected areas such as the Point Mugu wetlands area and Malibu beaches due to the ! 83! presence of flooded soils and tidal ranges, respectively. In most automatic ZSA situations though, “no data” was the cause for that result. Any area that did not have data was automatically restricted and, regardless of other factors, deemed ZSA. The areas of no data were mostly found along the northeast border of the SMMNRA with the San

Fernando Valley. It may be the case that these areas could be in fact suitable but due to lack of data, they were excluded from this analysis. If data could be acquired for these areas, it could provide important information about how northeast SMMNRA is affected along the San Fernando Valley, the most significant area of human impact (Swenson and

Franklin 2000).

Increased vegetation weighting in Model 2 showed a marked increase in HSAs

(+8%) and a similar decrease in HMSAs (-7%). This resulted from the upgrading of

HMSAs into HSAs based solely on increased vegetation weighting. Soil remained the same in all models and as such, did not create variations in the models. Border was minimally influential as it only affected a 500m interior portion of the SMMNRA boundary. For these areas, this MSA border value decreased the potential suitability area classification. Also, as shown in Model 3, increased slope weighting caused an increase in fragmentation of HSAs but not much change in total HSA PLAND values.

On all three models, HSAs were mostly surrounded by HMSAs, indicating a smooth transition for CHL movement (Brattstrom 1997). This was especially true for

Models 1 and 3. Model 2 though experienced an increase in HSA-MSA adjacency.

HMSAs in Model 2 declined, partly due to upgrading to HSAs but also due to downgrading to MSAs. Since the border weight for Models 2 and 3 were the same and the slope weight for Model 2 was lower than both Models 1 and 3, the only factor that

! 84! could have led to an increase in HSA-MSA adjacency was the increase in vegetation weighting. This was unexpected since Model 2 also showed a significant increase in total

HSA, roughly 8% higher than Models 1 and 3. Combined with the downgrading of

HMSAs into MSAs along HSA boundaries, this could indicate more pronounced negative changes in vegetation along HSA boundaries, possibly due to human influence. Since two of the three HSAs in Model 2 are part of protected state park land (Northern and Southern

Malibu Creek State Park), this could indicate that protection within these areas helps maintain HSA/HMSA conditions for CHLs.

Table 11 – Weighting and HSA/HMSA Results Comparison

Suitability Model Suitability Model Suitability Model #1 #2 #3 Weighting Vegetation 30% Vegetation 40% Vegetation 30%

Soil 30% Soil 30% Soil 30%

Slope 25% Slope 20% Slope 30%

Border 15% Border 10% Border 10%

Percent HSA 2.95% 11.75% 2.65% (PLAND) PLADJ of HSA 29.97% 40.21% 17.09%

Percent HMSA 66.27% 59.1% 66.1% (PLAND) PLADJ of 82.83% 76.01% 82.33% HMSA

! 85! 6.2 Research Question 1

Where are the three largest HSAs?

The answer to this research question varied depending on which model weighting was used. Overall, HSAs were identifiable on all three models and it was discernable which were the largest three for each. The largest HSAs did not have to be fully contiguous to qualify. Interspersion occurred in all of the largest HSAs but was mostly due to HMSAs. This provided nearly contiguous highly suitable areas by taking into account reasonable movement patterns (Munger 1984; Fair and Henke 1999; Wone and

Beauchamp 2003). Large HSAs (for this question) were not allowed to have intervening areas with a rating of MSA or below.

Suitability Model 1 had only 2.95% HSAs, almost the same percent as Model 3 but not in the same configuration. Model 1 showed Northern Malibu Creek State Park was the largest HSA, with Southern Malibu Creek State Park and Point Mugu State Park nearly equivalent in second largest area. Northern Malibu Creek State Park was the most contiguous HSA. All three HSAs appeared less fragmented on this model than on Models

2 and 3, even though the PLAND was similar to Model 3. Northern Malibu Creek State

Park HSA was consistently the least fragmented HSA on all three models.

Suitability Model 2 had 11.76% HSA, the largest percentage of any model. This increase in total HSA caused many smaller HSAs to form. Although total HSA increased, this did not create one very large HSA. The central HSA/HMSA complex though had a large increase in HSA interspersion. The largest HSAs were Northern Malibu Creek State

Park, Southern Malibu Creek State Park and Point Mugu State Park HSAs. Northern

Malibu Creek State Park was not always the absolute largest HSA on any given model

! 86! but was consistently the least fragmented.

Suitability Model 3 showed three more fragmented HSAs comprising 2.65% of the total area. These were Northern Malibu Creek State Park, Southern Malibu Creek

State Park and Rambla Pacifico. On this model, Point Mugu State Park only had a few fragmented sections of HSA and did not qualify. Although the HSA percentage was almost the same as Model 1, the HSA results for Model 3 were less dense and more fragmented than Model 1 due to increased slope weighting. This has potential implications that are further detailed in Research Question 3.

6.3 Research Question 2

Where are HSAs found in relation to park borders?

HSAs were found throughout SMMNRA on all three models but were mostly concentrated around Malibu Creek State Park, Point Mugu State Park and Rambla

Pacifico. There were smaller patches found, especially on Model 2, and may indicate potential border influence. Some of the larger HSAs fell near or within the 500-meter border zone. Northern Malibu Creek State Park and Point Mugu State Park were close enough to the park border impacted by the border buffer. These areas also formed part of

SMMNRA’s actual border, which exposed them directly to border-related human impact.

The Rambla Pacifico and southern Malibu Creek State Park HSAs were located deeper within SMMNRA and did not include or abut any of the border buffer zone. They remained unaffected by this study’s border classification but may be impacted by borders, roads, neighborhoods or other human impacts within the SMMNRA. As internal

! 87! human impact zones were not included in this research, no suitability impact can be determined for those aspects and that question remains a potential area of future study.

The northwest portion of northern Malibu Creek State Park was consistently

HMSA due its border proximity. If the border had not been considered in the weighting, the northern Malibu Creek State Park HSA would have extended to the park border. This

HSA was more fragmented in Model 3 than in either Model 1 or Model 2, though clearly not due to border weighting, as border was the same in Model 2 and 3. Northern Malibu

Creek State Park HSA was affected by the park border buffer zone as it formed part of the SMMNRA boundary.

Not accounted for in this study was the effect of the City of Calabasas within

SMMNRA. Although not within a designated “border” region, the presence of disturbed/urban/built-up vegetation and soil in Calabasas created a pocket of LMSA that directly bordered the northern Malibu Creek State Park HSA. This is a concerning situation with such highly suitable characteristics found so close to a known human impact area. More research is needed on this proximity and what effect it may have on

CHL populations within northern Malibu Creek State Park. Variations of this study could be done with border zones around the known cities within SMMNRA. This may also affect the results around Topanga and the coastal mountains of Malibu. This study did highlight areas of potential concern regarding human impact for future conservation planning.

The Point Mugu State Park HSA was located along the westernmost SMMNRA border boundary. On Model 2, that western HSA boundary was clearly cropped due to the border weighting. On Models 1 and 3, the effect was not noticeable as the Point Mugu

! 88! State Park HSA was minimal and fragmented on those overlays. Point Mugu State Park

HSA was affected by the park border buffer zone and, although protected, may still experience higher rates of human impact than those HSAs found further within

SMMNRA. This is similar to the northern Malibu Creek State Park results but the key difference is that most of the Point Mugu State Park HSA present in Model 2 was not near the border and did not have any cities infringing on it. The eastern portion of Point

Mugu State Park remains sheltered from future human development and does not border any cities. It also contains one of the largest non-fragmented HMSAs present on all models.

6.4 Research Question 3

Do HSAs appear to be fragmented in SMMNRA?

Each model visually showed different patterns of fragmentation and the metrics confirmed varied composition and configuration of SMMNRA habitat. It was clear though that, when looking only at HSAs, fragmentation was present between large HSAs and in human impact areas. That may have concerning implications such as decreased mate availability, increased invasive ant activity and/or disturbed vegetation (Goldberg

1983; Fair and Henke 1997; Beauchamp et al 1998; Suarez, Richmond and Case 2000;

Swenson and Franklin 2000; Martin and Murray 2010). HMSAs are also considered highly suitable habitats and fully functional as connecting segments between HSAs. If

HSA/HMSA complexes were considered though, combined PLAND values were on average 70% very suitable habitat. According to the California Wildlife Habitat

Relationships Data for CHL, the habitat size for population persistence is ideally about 2

! 89! square kilometers (greater than 500 acres). Above that, no extra benefit is conferred. The minimum size is 0.4 square kilometers (100 acres). These requirements also state that

75% of habitats present within that area must be suitable for CHLs (CDFG/CIWTG

2007). For this study, HSAs and HMSAs were considered very suitable. Although HSAs were fragmented in SMMNRA, the wide inclusion of HMSAs throughout the mountainous inner region created a relatively intact, very large area that is suitable on all accounts for CHLs.

On Model 1, HSA PLADJ was 29.97% and IJI was 12.18%. With the visual patterns on the model, this indicated low-moderate fragmentation of HSAs but this was balanced by the fact most of the interspersion was by HMSAs. On Model 2, HSA PLADJ was 40.21% and IJI was 16.23%. With the visual patterns on the model, this indicated a large increase in HSAs, again mixed with primarily HMSAs. This created a very large

HSA/HMSA complex in the core of the SMMNRA that was almost entirely non- fragmented, with HSAs evenly distributed throughout. The “Cotharin cluster” was an exception though, rated MSA or below due to low soil ratings. On Model 3, HSA PLADJ was 17.09% and IJI was 13.25%. Similar to Model 1, this indicated moderate fragmentation of large HSAs but each individual HSA was more fragmented.

Interspersion of HMSA with HSA was still the primary cause of HSA fragmentation.

Some HMSA-MSA fragmentation occurred in the “Cotharin Cluster,” areas impacted by cities, and severe slopes along a few canyons. Model 2 showed an increase of MSAs in central mountain range. Slope was the dominant reason for these MSA patches although the slope was still acceptable for migration through those areas. These intrusions did not prevent the connection of a substantial HSA/HMSA complex though.

! 90! The most severe locations of all types of fragmentation were the city of Calabasas and the

“Cotharin Cluster.” HMSA-MSA-LMSA inter-fragmentation appeared primarily in areas of human impact, not due to natural conditions. When examining specific HSAs,

Northern Malibu Creek State Park was not always the largest HSA on any of the models but was consistently the least fragmented. In terms of HSA/HMSA complexes, both Point

Mugu State Park and Topanga State Park (east to Sullivan Canyon) were consistently suitable habitat. These are very large areas: much larger than the ideal size necessary for population persistence (CDFG/CIWTG 2007). According to the data, an extensive network of partially connected HSAs exists due to most of the SMMNRA mountains designated HMSA. This suggests that CHL habitat is widely available throughout the

SMMNRA and well enough connected to allow for significant population exchange for mating needs (Goldberg 1983). It also means that CHL are more able to survive environmental pressures by staying within the core of the SMMNRA. Overall, the fragmentation of HSA/HMSA complexes was minimal even though specific HSA fragmentation was higher. The results show that highly suitable contiguous habitat does exist throughout the area, even though the SMMNRA is not fully suitable for CHLs.

! 91! 7.0 – CONCLUSIONS

This project identified four main HSAs present within SMMNRA according to the habitat suitability classification system used. Over all three models, northern and southern Malibu Creek State Park remained classified as HSAs. When vegetation was weighted more heavily in Model 2, more areas were identified as HSA. The areas defined as Point Mugu State Park and Rambla Pacifico were also identified as HSAs independently on different overlays. Large swaths of HMSAs that are also highly suitable for CHL connected these HSAs as one larger HSA/HMSA complex. Most of SMMNRA is habitable for CHL with only a few isolated areas deemed ZSA. Most ZSAs were due to lack of data, rather than actual characteristics present. In addition, fragmentation within the HSA/HMSA complex within the SMMNRA is minimal, but does occur in areas of high human impact.

7.1 Limitations

This study has limitations such as the lack of field observations and assumptions regarding the classification of habitat factors. The entire classification system is based on limited information available about habitat needs and suitability (Goldberg 1983; Fair and Henke 1997; Beauchamp et al 1998; Suarez, Richmond and Case 2000; Swenson and

Franklin 2000; Hernandez et al 2006; Martin and Murray 2010). Some information and assumptions regarding urban interface, agricultural impact, and home ranges was taken from studies on Texas Horned Lizards. It was difficult to classify soil series since no previous studies have been conducted regarding CHL suitability of specific soils found

! 92! within SMMNRA although the CWHR analysis did provide a framework for that classification. Classification was made based on comparing the information that was available about soil suitability in Coast Horned Lizards and other horned lizard species to the soil series descriptions provided by the USDA/NRCS soil survey (USDA/NRDC

2007). Vegetation data may not reflect actual condition on the CHL level (Dennison and

Roberts 2003). Analysis based on GIS data is subject to the accuracy of the data. GIS data inherently has accuracy issues due to availability, scale, resolution, and intended usage. The GIS-based soil series data used in this study was intended for larger scale analysis than how it was applied in this case but it was the only soil type data available.

Also, soil and vegetation data may not have been ground-truth checked for accuracy at an appropriate scale or at all (Syphard et al 2011). Actual CHL presence was not confirmed by field surveys. Field surveys may be needed to confirm actual habitat suitability on the ground. Graphics for qualitative comparisons were used to aid interpretation and explanation whenever possible. Since this study does not involve sampling to estimate

Coast Horned Lizard population to confirm their occupancy and absences, care must be taken in interpreting the conclusions of this study. This research was meant only as baseline habitat suitability models that suggest likely high suitability areas for Coast

Horned Lizards. This study assumes that no other conditions affect the habitat suitability of CHL and that the classifications are correct given the information provided, that the data used is still accurate and that those areas deemed suitable are in fact still suitable today. These classifications/models could be verified or updated in a number of ways such as ground-truth surveys, remote sensing data, or field surveys of CHL inhabitation.

Further research into additional suitability needs for CHL would also help expand the

! 93! usefulness of these models.

7.2 Impact on Future Research

These results may lead to greater protection of connecting areas that ensure the wide geographic range needed for optimal genetic diversity. The results of this research could be used to prioritize the conservation of HSA/HMSAs, provide baseline locations for future CHL field surveys, or even a starting point for assessing the habitat suitability of other ecologically related species like harvester ants. Habitat modeling may also aid in restoration of harvester ants to invasive-species-affected areas (Burger et al 2003;

Longcore 2003). Future conservation or development plans in the SMMNRA region should include the affects of human impact on the Coast Horned Lizard (Shearer et al

2006).

This study could also assist in evaluating the inclusion of the proposed Rim of the

Valley Corridor extensions into the national park system as a separate park unit or as an addition to SMMNRA by providing maps of potentially suitable CHL habitat in the

SMMNRA, as well as future studies that could look at CHL suitability in the proposed new areas. There are multiple proposed expansion plans and CHL habitat suitability modeling could be applied in the same manner as this research to any of those plans, provided there is data available for those areas. This Special Resource Study is examining the feasibility of expanding SMMNRA or creating a new park to include the Santa

Susana Mountains, Simi Hills, Conejo Mountains, Verdugo Mountains and/or Arroyo

Seco (National Park Service 2012). These areas are being evaluated for national significance in both ecological and cultural resources. These proposed boundary

! 94! adjustments could create linkages between known habitats and connect the areas of

SMMNRA, the Los Padres National Forest, the Angeles National Forest, and state and local habitat areas.

Part of the Rim of the Valley feasibility study includes a National Environmental

Policy Act (NEPA) assessment of these areas (National Park Service 2012). As the Coast

Horned Lizard is a species of special concern (SSC), the results of this research may be of great use as part of that environmental assessment. This map will be shared with the

National Park Service and interested governmental agencies and NGOs interested in the survival of Coast Horned Lizards. The Coast Horned Lizard is a valuable part of the ecosystem in southern California as one of the few ant-eating species present. The creation of a new adjoining park or an extension of SMMNRA could greatly increase the survival of Coast Horned Lizards by expanding protected habitats and limiting future human impact. As this research showed, human impact is the largest force in fragmentation of suitable habitats for CHL in SMMNRA. Habitat loss and fragmentation due to human impact remain key concerns for National Park System planners. According to the NPS “Rim of the Valley Corridor Special Resource Study, Newsletter #3,” without functional landscape connections for migration, dispersal, and other ecological functions, some native species in the Santa Monica Mountains may cease to exist there in the future

(National Park Service 2012). The survival of the Coast Horned Lizard in the SMMNRA and surrounding areas depends not just on knowing their habitat suitability characteristics but having a geographic model of their most suitable habitats in these areas. This research presented here provides that necessary and desired information for any future conservation planning.

! 95! REFERENCES

Andelman, Sandy J. 2011. "Conservation Science Outside the Comfort Zone.” Nature

475: 291.

Beauchamp, Beate, Bernard Wone, Michael Kutilek, and Shannon Bros. 1998. "Habitat

Use of the Flat-tailed Horned Lizard (Phrynosoma Mcallii) in a Disturbed

Environment." Journal of Herpetology 32(2): 210-216.

Bailey, H.P. 1966. Climate of Southern California. Berkeley, CA: University of

California Press.

Beier, P. 1993. “Determining Minimum Habitat Areas and Habitat Corridors for

Cougars.” Conservation Biology 7: 94–108.

Bolger, Douglas T. 2007. “Spatial and Temporal Variation in the Argentine Ant Edge

Effect: Implications for the Mechanism of Edge Limitation” Biological Conservation

136: 295-305.

Bowler, Peter A. 2000. “Ecological Restoration of Coastal Sage Scrub and its Potential

Role in Habitat Conservation Plans.” Ecological Management 26: S85- S96.

Brattstrom, Bayard H. 1997. "Status of the Subspecies of the Coast Horned Lizard,

Phrynosoma coronatum." Journal of Herpetology 31(3): 434-436.

Brooks, Thomas M., Russell A. Mittermeier, Christina G. Mittermeier, Gustavo A.B. da

Fonseca, Anthony B. Rylands, William R. Konstant, Penny Flick, John Pilgrim, Sara

Oldfield, Geogina Magin, and Craig Hilton-Taylors. 2002. “Habitat Loss and

Extinction in the Hotspots of Biodiversity.” Conservation Biology 16 (4) 909- 923.

! 96! Brussard, Peter F. 1991. “The Role of Ecology in Biological Conservation.” Ecological

Applications 1(1): 6-12.

Burger, J.C., R.A. Redak, E.B. Allen, J.T. Rotenberry and M.F. Allen. 2003. “Restoring

Arthropod Communities in Coastal Sage Scrub.” Conservation Biology 17(2): 460-

467.

Burrow, Anna, Richard Kazmaier, Eric Hellgren, and Donald Ruthven III. 2001.

"Microhabitat Selection by Texas Horned Lizards in Southern Texas." Journal of

Wildlife Management 65(4): 645-652.

California Department of Fish and Game and California Interagency Wildlife Task Group

(CDFG/CIWTG). 2007. “Species Notes for Coast Horned Lizard (Phrynosoma

coronatum): California Wildlife Habitat Relationships (CWHR) System Level II

Model Prototype.”

Cassidy, Kelly M. and Christian Gore. 2000. “The Role of Private and Public Lands in

Conservation of At-Risk Vertebrates in Washington State.” Wildlife Society Bulletin

28(4): 1060-1076.

Clarke B. 2002. “Current Legislated Boundary for the Santa Monica Mountains National

Recreation Area. National Park Service - Santa Monica Mountains National

Recreation Area.” Geospatial Dataset-1042878.

Cody, M.L. 1986. “Diversity, Rarity and Conservation in Mediterranean Climate

Regions.” In M.E. Soule (eds.) Conservation Biology. Sunderland, MA, Sinauer

Associates: 122-152.

! 97! Comrack, L., B. Bolster, J. Gustafson, D. Steele and E. Burkett. 2008. “Species of

Special Concern: A Brief Description of an Important California Department of Fish

and Game Designation.” California Department of Fish and Game, Wildlife Branch,

Nongame Wildlife Program Report.

Cooper, William E., Jr, and Wade Sherbrooke. 2009. "Prey Chemical Discrimination by

Tongue Flicking is Absent in the Texas Horned Lizard, Phrynosoma Cornutum."

Journal of Herpetology 43(4): 688-692.

Cunningham, John D. 1955. “Arboreal Habits of Certain Reptiles and Amphibians in

Southern California.” Herpetologica. 11(3): 217-220.

Davis, Frank W., Peter A. Stine, and David M. Stoms. 1994. “Distribution and

Conservation Status of Coastal Sage Scrub in Southwest California.” Journal of

Vegetation Science 5: 743-756

Dennison, Philip E. and Dar. A. Roberts. 2003. “The Effects of Vegetation Phenology on

Endmember Selection and Species Mapping in Southern California Chaparral.”

Remote Sensing of Environment 87: 295–309. (Dennison and Roberts 2003)

Donaldson, W., A.H. Price. and J. Morse. 1994. “The Current Status and Future

Prospects of the Texas Horned Lizard (Phrynosoma cornutum) in Texas.” Texas

Journal of Science 46(2): 97-113.

Dratch, Peter A., and Loyal A. Mehrhoff. 2002. "Endangered Species and the National

Park Service." Endangered Species Update 19(2): 1-4.

Endriss, Debora, Eric Hellgren, Stanley Fox, and Raynond Moody. 2007. "Demography

of an Urban Population of the Texas Horned Lizard (Phrynosoma cornutum) in

Central Oklahoma." Herpetologica 63(3): 320-331.

! 98! ESRI. 2012. ArcGIS (ArcMap, Version 10.1)

Fahrig, Lenore. 1997. "Relative Effects of Habitat Loss and Fragmentation on Population

Extinction." The Journal of Wildlife Management 61(3): 603-610.

Fair, W. Scott, and Scott Henke. 1997. "Effects of Habitat Manipulations on Texas

Horned Lizards and Their Prey." The Journal of Wildlife Management 61(4): 1366-

1370.

---. 1999. "Movements, Home Ranges, and Survival of Texas Horned Lizards

(Phrynosoma cornutum)." Journal of Herpetology 33(4): 517-525.

Fisher, Robert, Andrew Suarez, and Ted Case. 2002. "Spatial Patterns in the Abundance

of the Coastal Horned Lizard." Conservation Biology 16(1): 205-215.

Goldberg, Stephen R. 1983. "Reproduction of the Coast Horned Lizard, Phrynosoma

coronatum, in Southern California." The Southwestern Naturalist, 28(4): 478-479.

Goldstein, Paul, Yael Wyner, Phaedra Doukakis, Mary G. Egan, George Amato, Howard

Rosenbaum, and Rob DeSalle. 2005. “Theory and Methods for Diagnosing Species

and Populations in Conservation.” Annals of the Missouri Botanical Garden 92(1):

12-27

Grant, Tyler J. and Paul F. Doherty Jr. 2007. “Monitoring of the Flat-Tailed Horned

Lizard With Methods Incorporating Detection Probability.” Journal of Wildlife

Management 71(4): 1050-1056.

Hargis, Christina D., John A. Bissonette, and John L. David. 1998. “The Behavior of

Landscape Metrics Commonly Used in the Study of Habitat Fragmentation.”

Landscape Ecology 13:167–186.

! 99! Hellgren, Eric C., Anna L.Burrow, Richard T Kazmaier, and Donald C. Ruthven. 2010.

“The Effects of Winter Burning and Grazing on Resources and Survival of Texas

Horned Lizards in a Thornscrub Ecosystem.” Journal of Wildlife Management 74(2):

300-309.

Henke, Scott. 2003. "Baseline Survey of Texas Horned Lizards, Phrynosoma cornutum,

in Texas." The Southwestern Naturalist 48(2): 278-282.

Hernandez, Pilar A, Catherine H Graham, Lawrence L Master, and Deborah L Albert.

2006. "The Effect of Sample Size and Species Characteristics on Performance of

Different Species Distribution Modeling Methods." Ecography, 29(5): 773-785.

Hodges, Wendy L. and Kelly R. Zamudio. 2004. “Horned Lizard (Phrynosoma)

Phylogeny Inferred from Mitochondrial Genes and Morphological Characters:

Understanding Conflicts Using Multiple Approaches.” Molecular Phylogenetics and

Evolution 31: 961–971.

Holway, David A. 2005. “Edge Effects of an Invasive Species Across a Natural

Ecological Boundary.” Biological Conservation 121: 561–567.

Holway, David A. and Andrew V. Suarez. 2006. “Homogenization of Ant Communities

in Mediterranean California: The Effects of Urbanization and Invasion.” Biological

Conservation 127:319-326.

Huffman, M. 1998. Wild Heart of Los Angeles: The Santa Monica Mountains. Dublin,

Ireland, Robert Rinehart Press.

Jennings, Mark R. 1987. “Impact of the Curio Trade for San Diego Horned Lizards

(Phrynosoma coronatum blainvillii) in the Los Angeles Basin, California: 1885-

1930” Journal of Herpetology 21(4): 356-358.

! 100! Kowarik, Ingo. 2011. "Novel Urban Ecosystems, Biodiversity, and Conservation.”

Environmental Pollution 159: 1974-1983.

Lahti, Megan, and Daniel Beck. 2008. "Ecology and Ontogenetic Variation of Diet in the

Pigmy Short-Horned Lizard (Phrynosoma douglasii)." The American Midland

Naturalist 159(2): 327-339.

Leaché, Adam, Michelle Koo, and Carol Spencera. 2009. "Quantifying Ecological,

Morphological, and Genetic Variation to Delimit Species in the Coast Horned Lizard

Species Complex (Phrynosoma)." Proceedings of the National Academy of Sciences

of the United States of America 106(30): 12418-12423.

Li, H., J. F. Franklin, Swanson, F. J. and Spies, T.A. 1993. “Developing Alternative

Forest Cutting Patterns: A Simulation Approach.” Landscape Ecology 8: 63–75

Longcore, Travis. 2003. “Terrestrial Arthropods as Indicators of Ecological Restoration

Success in Coastal Sage Scrub (California, U.S.A.).” Restoration Ecology 11(4):

397–409

Lynn, Robert T. 1965. “A Comparative Study of Display Behavior in Phrynosoma

(Iguanidae).” The Southwestern Naturalist 10(1): 25-30.

MacMahon, James A., John F. Mull and Thomas O. Crist. 2000. “Harvester Ants

(Pogonomyrmex spp.): Their Community and Ecosystem Influences.” Annual Review

of Ecological Systematics 31: 265-291.

Martin, Leigh J. and Brad R. Murray. 2010. “A Predictive Framework and Review of the

Ecological Impacts of Exotic Plant Invasions on Reptiles and Amphibians.”

Biological Reviews 2011(86): 407-419.

! 101! Mathies, Tom, and Daniel Martin. 2008. "Overwintering Site Selection by Short-Horned

Lizards (Phrynosoma hernandesi) in Northeastern Colorado." Journal of Herpetology

42(1): 163-171.

Mayer, A.L., and G. N. Cameron. 2003. “Consideration of Grain and Extent in

Landscape Studies of Terrestrial Vertebrate Ecology.” Landscape and Urban

Planning 65 (4): 201-217

Meyers, J. J., A. Herrel and K.C. Nishikawa. 2006. “Morphological Correlates of Ant

Eating in Horned Lizards (Phrynosoma).” Biological Journal of the Linnean Society,

89: 13–24.

McGarigal, K., SA Cushman, and E Ene. 2012. FRAGSTATS v4: Spatial Pattern

Analysis Program for Categorical and Continuous Maps. Computer software program

produced by the authors at the University of Massachusetts, Amherst.

http://www.umass.edu/landeco/research/fragstats/fragstats.html Last Accessed:

2/23/2013.

Middendorf, George A. III, Wade C. Sherbrooke, and Eldon J. Braun. 2001.

“Comparison of Blood Squirted from the Circumorbital Sinus and Systemic Blood in

a Horned Lizard, Phrynosoma cornutum.” The Southwestern Naturalist 46(3): 384-

387.

Middendorf, George A. III, Wade C. Sherbrooke. 1992. “Canid Elicitation of Blood-

Squirting in a Horned Lizard (Phrynosoma cornutum).” Copeia 1992(2): 519-527.

Montanucci, Richard. 1989. "The Relationship of Morphology to Diet in the Horned

Lizard Genus Phrynosoma." Herpetologica 45(2): 208-216.

! 102! Moody, J.V., O.F. Frankie, and F.W. Merickel. 1981. “The Distribution of Fire Ants,

Solenopsis, in Western Texas (Hymenoptera, Formicidae).” Journal of the Kansas

Entomological Society 54: 469-480.

Munger, James. 1984. "Home Ranges of Horned Lizards (Phrynosoma): Circumscribed

and Exclusive?" Oecologia 62(3): 351-360.

Munger, James. 1984. "Long-Term Yield from Harvester Ant Colonies: Implications for

Horned Lizard Foraging Strategy." Ecology 65(4): 1077-1086.

National Park Service. 2006. “Vegetation Classification of the Santa Monica Mountains

National Recreation Area and Environs in Ventura and Los Angeles Counties,

California” Booklet.

National Park Service. 2012. Santa Monica Mountains National Recreation Area.

“Mediterranean Ecosystems.”

http://www.nps.gov/samo/naturescience/mediterraneanecosystem.htm. Last

Accessed: 2/23/2013.

National Park Service, 2012. Santa Monica Mountains National Recreation Area.

“Mammals.” http://www.nps.gov/samo/naturescience/mammals.htm. Last Accessed:

2/23/2013.

National Park Service. 2012. “Rim of the Valley Special Resource Study, Newsletter #3”

http://www.nps.gov/pwro/rimofthevalley. Last Accessed: 2/23/2013.

Newbold, T. A. Scott, and James MacMahon. 2009. "Spatial and Seasonal Dietary

Patterns of the Desert Horned Lizard (Phrynosoma platyrhinos): Harvester Ant

Specialist or Generalist Ant Feeder?" Canadian Journal of Zoology 87(2): 112-123.

! 103! Olden, Julian D. 2006. “Biotic Homogenization: A New Research Agenda for

Conservation Biogeography.” Journal of Biogeography 2006(33): 2027-2039. (Olden

2006)

Pianka, Eric, and William Parker. 1975. "Ecology of Horned Lizards: A Review with

Special Reference to Phrynosoma platyrhinos." Copeia 1975(1): 141-162.

Plantrich, R.F. 1990. “Monitoring of Golden Eagle Nests at Santa Monica Mountains

National Recreation Area - Examples of Resource Inventory and Monitoring in

National Parks of California.” U.S. Department of the Interior, Washington, D.C.,

USA. 49–55.

Powell, G. Lawrence, Anthony Russell, and Patrick Fargey. 1998. "The Distribution of

the Short-Horned Lizard Phrynosoma hernandezi in Saskatchewan, Canada."

Northwestern Naturalist 79.1: 19-26.

Price, A.H. 1990. Phrynosoma cornutum. Catalogue of American Amphibians and

Reptiles 469: 1-7.

Quinn, R.D. 1990. “Habitat Preferences and Distribution of Mammals in California

Chaparral.” Research paper PSW-202. Pacific Southwest Research Station, Forest

Service, U.S. Department of Agriculture, Berkeley, CA, USA.

Rahman, Abdullah F., John A. Gamon, Daniel A. Sims, and Miriam Schmidts. 2003.

“Optimum Pixel Size for Hyperspectral Studies of Ecosystem Function in Southern

California Chaparral and Grassland.” Remote Sensing of Environment 84: 192–207.

Reeder, Tod W. and John J. Wiens. 1996. “Evolution of the Lizard Family

Phrynosomatidae as Inferred form Diverse Types of Data.” Herpetological

Monographs 10: 43-84.

! 104! Roberts, D.A, M Gardner, R Church, S Ustin, G Scheer, and R.O Green. 1998. "Mapping

Chaparral in the Santa Monica Mountains Using Multiple Endmember Spectral

Mixture Models." Remote Sensing of Environment, 65(3): 267-279.

Rose, Barbara. 1982. “Lizard Home Ranges: Methodology and Functions.” Journal of

Herpetology 16(3): 253-269.

Shearer, Allan W, David A Mouat, Scott D Bassett, Michael W Binford, Craig W

Johnson, and Justin A Saarinen. 2006. "Examining Development-related

Uncertainties for Environmental Management: Strategic Planning Scenarios in

Southern California." Landscape and Urban Planning, 77(4): 359-381.

Sherbrooke, Wade, and George A. Middendorf III. 2001. "Blood-squirting Variability in

Horned Lizards (Phrynosoma)." Copeia 2001(4): 1114-1122.

Sherbrooke, Wade. 2003. Introduction to Horned Lizards of North America. Berkeley:

California Natural History Guides.

---. 2004. “Responses of Kit Foxes (Vulpes macrotis) to Antipredator Blood-Squirting

and Blood of Texas Horned Lizards (Phrynosoma cornutum)” Copeia 2004(3): 652-

658.

---. 2008. "Antipredator Responses by Texas Horned Lizards to Two Snake Taxa with

Different Foraging and Subjugation Strategies.” Journal of Herpetology 42(1): 145-

152.

Soltz, Roberta Lynn. 1984. “Time and Energy Budgets of the Red-Tailed Hawk in

Southern California.” The Southwestern Naturalist 29(2):149-156.

! 105! Stark, Richard, Stanley Fox, and David M. Leslie Jr. 2005. "Male Texas Horned Lizards

Increase Daily Movements and Area Covered in Spring: A Mate Searching Strategy?"

Journal of Herpetology 39(1): 169-173.

Suarez, Andrew, and Ted Case. 2002. "Bottom-Up Effects on Persistence of a Specialist

Predator: Ant Invasions and Horned Lizards." Ecological Applications 12(1): 291-

298.

Suarez, Andrew, Jon Richmond, and Ted Case. 2000. "Prey Selection in Horned Lizards

Following the Invasion of Argentine Ants in Southern California." Ecological

Applications 10(3): 711-725.

Swenson, Jennifer J., and Janet Franklin. 2000. “The Effects of Future Urban

Development on Habitat Fragmentation in the Santa Monica Mountains.” Landscape

Ecology 15(8): 713-730.

Syphard, Alexandra A., Keith C. Clarke, Janet Franklin, Helen M. Regan, Mark

Mcginnis. 2011. “Forecasts of Habitat Loss and Fragmentation due to Urban Growth

are Sensitive to Source of Input Data.” Journal of Environmental Management 92:

1882-1893.

Thomas, Jack Ward, and Daniel H. Pletcher. 2000. "The Convergence of Ecology,

Conservation Biology, and Wildlife Biology: Necessary or Redundant?" Wildlife

Society Bulletin 28(3): 546-549.

United States Department of Agriculture, Natural Resources Conservation Service. 2006.

“Soil Survey of Santa Monica Mountains National Recreation Area, California.”

Accessible online at: http://soils.usda.gov/survey/printed_surveys/. Last Accessed:

2/23/2013.

! 106! Wiens, J. A. 1989. “Spatial Scaling in Ecology.” Ecology Functional 3 (4): 385-397.

Wiernasz, Diane C., and Blaine J. Cole. 1995. "Spatial Distribution of Pogonomyrmex

occidentalis: Recruitment, Mortality and Overdispersion.” Journal of Animal Ecology

64(4): 519-527.

Whitford, Walter, and Martha Bryant. 1979. "Behavior of a Predator and Its Prey: The

Horned Lizard (Phrynosoma cornutum) and Harvester Ants (Pogonomyrmex spp.)."

Ecology 60(4): 686-694.

Whiting, Martin, James Dixon, and Roy Murray. 1993. "Spatial Distribution of a

Population of Texas Horned Lizards (Phrynosoma cornutum: Phrynosomatidae)

Relative to Habitat and Prey." The Southwestern Naturalist 38(2): 150-154.

Wilcox, B. A. and Murphy, D. D. 1985. “Conservation Strategy: The Effects of

Fragmentation on Extinction.” American Naturalist 125: 879–887

Wone, Bernard, and Beate Beauchamp. 2003. "Movement, Home Range, and Activity

Patterns of the Horned Lizard, Phrynosoma mcallii." Journal of Herpetology 37(4):

679-686.

Xu, Yan. 2001. “Spatial Patterns of Population Growth and Habitat Change in the Santa

Monica Mountains Under Different Local Policy Constraints.” University of Southern

California, ProQuest, UMI Dissertations Publishing.

Yetter, Laura. 2012. “Landscape Pattern: A Determinate of Coastal California

Gnatcatcher (Polioptila californica californica) Occupation.” Thesis. California State

University, Northridge.

Young, Kevin, Edmund Brodle, and Edmund Brodle III. 2004. "How the Horned Lizard

Got Its Horns." Science 304(5667): 65.

! 107! Zeiner, D.C., W.F. Laudenslayer, Jr., K.E. Mayer, and M. White, eds, revised by J

Vindum. 1988-1990, rev 1997. “California's Wildlife. Vol. I-III.” California

Department of Fish and Game, Sacramento, California.

! 108! APPENDIX A

Species Notes for Coast Horned Lizard (Phrynosoma coronatum): California Wildlife Habitat Relationships (CWHR) System Level II Model Prototype

California Department of Fish and Game California Interagency Wildlife Task Group

(CDFG/CIWTG 2007)

This document is part of the California Wildlife Habitat Relationships (CWHR) System, operated and maintained by the California Department of Fish and Game (CDFG) in cooperation with the California Interagency Wildlife Task Group (CIWTG). The information will be useful for environmental assessments and wildlife habitat management.

Notes such as these were prepared for 32 species by the US Forest Service Pacific Southwest Research Station as part of a 2000/2001 contract with CDFG. Each is part of a prototypical “Level II” model for a species. As compared with the “Level I” or matrix models initially available in the CWHR System, “Level II” models incorporate spatial issues such as size of a habitat patch and distance between suitable habitat patches.

The notes are divided into three major sections. First, “Distribution, Seasonality and Habitats” represents information in the existing Geographic Information System (GIS) range data and in the Level I matrix model for a species. There is a vector- based GIS layer of geographic range and seasonality for each species in CWHR as well as a matrix containing all suitability ratings – High (H), Medium (M), Low (L) or Unsuitable (-) – by habitat (e.g. BOW or Blue Oak Woodland), stage (e.g. 4P or small tree, open canopy) and life requisite (reproduction, cover, or feeding.).

Second, “Required Attributes of Suitable Habitat Patches” represents spatially- explicit requirements of a species. The information here builds upon what is known about habitat patch size and the most critical attributes of a habitat patch needed by an individual of the species.

Third, “Spatial Habitat Requirements for Persistence of Population” represents estimates of the amount of habitat needed to maintain a population of a species. This may be considered the starting point for a “Level III” CWHR model, which would take into account spatial issues as well as a number of population parameters not yet incorporated into CWHR. Such information is included for most, but not all, Level II- modeled species.

! 109! Model Parameter Threshold Value(s) for Species

Suitable Habitats Species finds suitability (H --->L) for reproduction, cover and/or feeding in some Habitats rated in the California Wildlife or all stages of: Alkali Desert Scrub, Annual Habitat Relationships (CWHR) System as Grassland, Blue Oak Woodland, Blue Oak – high (H), medium (M), or low (L) suitability Foothill Pine, Chamise-Redshank for reproduction, cover, or feeding Chaparral, Closed-cone Pine Cypress, Coastal Oak Woodland, Coastal Scrub, Dryland Grain Crops, Eucalyptus, Irrigated Row and Field Crops, Juniper, Mixed Chaparral, Montane Hardwood, Perennial Grassland, Ponderosa Pine, Rice, Valley Foothill Riparian, Valley Oak Woodland, and Vineyard.

Water Water is irrelevant for suitability. Species does not require permanent water. Whether water is required, enhances, or is irrelevant for habitat suitability

Patch Size 0.5 acre (L)

L = low suitability. This is the minimum 10 acres (H) patch size for persistence of an individual.

H = high suitability. Above this patch size, area alone does not increase habitat suitability for an individual.

Edges Edges are not required by this species.

Requirements for a transition between two life form types Structural Habitat Attributes Friable soils, especially sandy soils, are Requirements for live vegetation, dead or essential for reproduction and cover decadent vegetation, vegetation residues, (burrows placed in friable soils). Species physical features, or human-made features needs anthills with logs or rocks for basking close to either soils suitable for burrowing or burrows covered by logs or rocks.

Food Species eats terrestrial insects, especially ants.

Spatial Habitat Requirements for Persistence of Population

Lowest suitability = 100 acres, if suitable patches cover at least 75% of area, are of a minimum size (see above), and are a maximum of 15 meters apart

Highest suitability = greater than 500 acres, if suitable patches cover at least 75% of area, are of a minimum size (see above), and are less than 5 meters apart

! 110! APPENDIX B

Coast Horned Lizard Range Map for California

! 111! APPENDIX C

Class Level Metrics for the Suitability Models

TYPE PLAND AREA_AM ENN_AM PLADJ IJI

MODEL 1 ZSA 4.2936 554.3549 257.3207 71.109 63.7575 MODEL 1 LSA 0.0161 1.2 3151.0068 0 41.3718 MODEL 1 LMSA 3.3591 45.4728 298.1434 42.5757 48.086 MODEL 1 MSA 23.1082 676.7403 208.9997 55.7139 33.5137 MODEL 1 HMSA 66.269 35268.5046 200.756 82.8303 45.8265 MODEL 1 HSA 2.9539 53.3486 266.6996 29.9727 12.1779

MODEL 2 ZSA 4.2936 554.3549 257.3207 71.109 73.1539 MODEL 2 LSA 0.3212 36.3668 519.687 52.2613 54.8118 MODEL 2 LMSA 5.9126 117.2943 251.3657 50.6416 59.775 MODEL 2 MSA 18.616 270.507 212.0794 54.6649 49.2525 MODEL 2 HMSA 59.0974 29293.1335 201.4311 76.0188 60.6245 MODEL 2 HSA 11.7591 71.5019 210.1708 40.2128 16.2339

MODEL 3 ZSA 4.1575 585.4951 241.6921 72.662 64.8065 MODEL 3 LSA 0.0403 3.08 1551.0527 18 45.6402 MODEL 3 LMSA 3.425 43.5831 273.8319 40.0612 49.2433 MODEL 3 MSA 23.6251 691.8823 207.783 55.5859 33.4471 MODEL 3 HMSA 66.0999 34667.6968 200.5404 82.3294 46.2756 MODEL 3 HSA 2.6523 7.8358 288.1471 17.0925 13.2473

! 112!