Species Status Assessment for Graptopetalum Bartramii

Species Status Assessment Report for

Graptopetalum bartramii Bartram’s stonecrop

April 2018 Southwest Region United States Fish and Wildlife Service Arizona Ecological Services Office Tucson, Arizona

This document was prepared by Julie Crawford, Arizona Ecological Services Office and Hayley Dikeman, Southwest Regional Office. We thank Angela Anders, Southwest Regional Office, for reviews of drafts. Valuable input was provided by George Ferguson, University of Arizona; Steve Buckley, National Park Service; George Montgomery, formerly of the Arizona Sonora Desert Museum; Jessica Simms, National Park Service; Angela Barclay, National Forest Service, and Kathryn Kennedy, National Forest Service.

Suggested reference:

United States Fish and Wildlife Service. 2018. Species status assessment report for Graptopetalum bartramii (Bartram’s stonecrop). United States Fish and Wildlife Service Arizona Ecological Services Office, Tucson, Arizona. 117 pp. + 2 appendices.

iii

Species Status Assessment Report for Graptopetalum bartramii Bartram’s stonecrop Prepared by the United States Fish and Wildlife Service

EXECUTIVE SUMMARY

This species status assessment reports the results of the comprehensive status review for Graptopetalum bartramii (Bartram’s stonecrop) through an analytical approach assessing the species needs, current status, and future status of a species using the best available information, and provides a thorough account of the species’ overall viability. For the purpose of this assessment, we generally define viability as the ability of G. bartramii to sustain populations in natural systems over time. The SSA Framework uses the conservation biology principles of resiliency, redundancy, and representation (collectively known as the “3Rs”) as a lens to evaluate the current and future condition of the species.

Graptopetalum bartramii is a succulent of the Crassulaceae or stonecrop family. The species typically occurs on rocky outcrops in deep, narrow canyons in heavy cover of litter and shade; and typically within 10 meters (m; 32.8 feet (ft)) of streambeds, springs, or seeps. It is known to have historically occurred in 33 separate populations within 13 isolated sky island mountain ranges, 10 in southern Arizona and 3 in northern Mexico. We are aware of four populations that have become extirpated in the United States in recent years, and a fifth, which has contracted in size. In three instances, extirpation was associated with the drying of habitat, which rendered it no longer suitable for the species to persist; we do not know the cause of extirpation in the fourth instance. Currently, there are 29 extant populations across 12 mountain ranges in the United States and Mexico. We are aware of 3,726 adult individuals across the entire range within the United States, including an assumed 10 plants from one U.S. population that has not been revisited, but presumed to be small (Thomas Canyon). Similarly, the three populations in Mexico have not been revisited and no counts ever made at these sites. We assume a small number of plants (10) at each of these locations, for a total of 3,756 across the entire range of the species. We also assume habitat, where these populations grow, to be in Low condition.

Graptopetalum bartramii needs multiple resilient populations. Resilient populations are those able to withstand stochastic events arising from spatially and temporally random factors, and that are distributed widely across its range, to maintain its persistence into the future and to avoid extinction. Several factors influence the resiliency of a population in response to stochastic events. These factors are: • Abundance – populations large enough that local stochastic events do not eliminate all individuals, allowing the overall population to recover from any one event • Subpopulations – multiple subpopulations per population so that local stochastic events do not eliminate the entire population • Recruitment – the number of seedlings exceeds the number of dead or dying individuals • Riparian elements – typically located within 10 m (32.8 ft) of a water source, which is presumed to increase humidity

• Precipitation – to maintain soil moisture, cooler temperatures, and humidity in the microenvironment • Shade – litter and deep shade from rock walls and / or overstory vegetation which is presumed to increase humidity • Substrate – crevices in solid bedrock or in shallow soil pockets • Pollinators – sufficient to ensure seed production, as the species does not reproduce vegetatively

To assess the population resiliency levels for current condition, we used the first six population and habitat factors listed above, as they are the primary factors influencing G. bartramii. These include all of the above factors except substrate and pollinators. We did not include these last two because they do not appear to be limiting factors for this species. For each of the six population and habitat factors we developed condition categories (High, Moderate, Low, and Extirpated) to assess the condition of each factor for each population, in order to determine overall population resiliency. Tables ES–1 and ES-2 provide our assessment of the current and future conditions of G. bartramii populations. To assess current condition of the species, we considered representation and redundancy. Representation is the ability to adapt to changing environmental conditions as measured by the breadth of genetic or environmental diversity within and among populations. Redundancy is the ability of a species to withstand catastrophic events, measured by the number of populations, their resiliency, and their distribution and connectivity. Genetic analysis of this species has not been conducted within or among populations or sky islands. However, sky island populations on different mountain ranges are widely separated (ranging from roughly 14 to 42 km (8.7 to 26 mi) apart), making cross- pollination highly unlikely, and most of the populations contain small numbers of individuals. Currently, there are 29 populations across 12 mountain ranges, with lack of connectivity among mountain ranges.

To evaluate the biological status of G. bartramii into the future, we assessed a range of conditions over 10 and 40 years to allow us to consider the species’ resiliency, redundancy, and representation. Our analysis of the past, current, and future influences on what G. bartramii needs for long term viability revealed that there are a number of stressors to this species including two primary stressors, related to habitat changes, which pose the greatest risk to future viability of the species. A stressor is a chemical or biological agent, environmental condition, external stimulus or an event that causes stress to an organism, and risk is the possibility of the stressor impacting the organism. These stressors include 1) groundwater extraction and drought that may reduce nearby water levels and humidity within G. bartramii habitat, and 2) altered fire regimes leading to erosion of G. bartramii habitat, sedimentation that could cover individuals, and loss of overstory shade trees. These stressors play a large role in the future viability of G. bartramii, especially for smaller populations. If populations lose resiliency, they are more vulnerable to extirpation, with resulting losses in representation and redundancy. Each population faces varying levels of risk into the future from natural and anthropogenic stressors including the following:

• Loss of water in nearby drainages from mining and drought • Erosion, sedimentation, and burial from mining, livestock, wildlife, recreation trails and roads, cross border violators, and post-wildfire runoff

• Trampling from humans, wildlife, and livestock • High severity wildfires ignited from recreationists, cross border violators, and lightning • Loss of shade from mining, drought, insect predation, flooding, and wildfire • Higher frequencies of freezing and flooding events from current and future climate change • Loss of seedling, immature, and adult plants, and reproduction from current and future drought • Predation of individuals and shade trees • Illegal collection

Given our uncertainty regarding the species’ persistence in Mexico and the future of nearby water and shade that create a humid microhabitat within populations, we have forecasted what G. bartramii may have in terms of resiliency, redundancy, and representation under four future plausible scenarios (Table ES–1 and 2).

Under scenario 1 – Continuation - We would expect the viability of G. bartramii to be characterized by a loss of resiliency, representation, and redundancy at the level that is currently occurring. At the 10-year time step, no populations would be in High condition, 4 populations (12 percent) would be in Moderate condition, 23 populations (70 percent) would be in Low condition and more susceptible to loss, and 6 populations (18 percent) would be extirpated. Within Scenario 1 we assume impacts from drought, climate change, and other stressors continue as in the near past. We think this is highly likely to occur within the 10-year time step with decreasing likelihood at future timesteps. This expectation is based on climate change projections portraying emissions increases, resulting in increased impacts to the species.

Under scenario 2 – Conservation – We would expect the viability of G. bartramii to be characterized by higher levels of resiliency, representation, and redundancy than it exhibits under the current condition. However, because current stressors remain in place, conservation measures would result in populations similar to those under current conditions at best: static at the 40-year time step, with no populations in High condition, 11 populations in Moderate condition (33 percent), 18 in Low condition (55 percent), and the 4 populations (12 percent) currently extirpated remain extirpated. Within Scenario 2, we assume impacts from drought, climate change, and other stressors continue as in Scenario 1, but with conservation measures addressing some stressors to the species such as nonnative plant control, forest thinning, and prevention of human caused wildfire. However, because climate change impacts are projected to increase and no conservation measures could address drying of habitat, we think this Scenario is somewhat likely in the long term.

Under scenario 3 – Moderate Effects, we would expect the viability of G. bartramii to be characterized by lower levels of resiliency, representation, and redundancy than it has in the Continuation scenario. At the 40-year time step, no populations would be in High condition, 4 populations (12 percent) would remain in Moderate condition, 17 populations (52 percent) would be in Low condition, and 12 populations (36 percent) would be extirpated. Within Scenario 3 we assume increased confidence in climate change impacts based on IPCC projections, in which they state high confidence in emissions scenario 4.5 being exceeded. Based on IPCC’s projections, we think that in the long term this is moderately likely.

Under scenario 4 – Major Effects, we would expect the viability of G. bartramii to be characterized by lower levels of resiliency, representation, and redundancy than under the Moderate effects scenario. At the 40-year time step, no populations would be in High condition, 1 population (3 percent) would be in Moderate condition, 11 (33 percent) would be in Low condition and more susceptible to loss, and 21 (64 percent) would be extirpated. For Scenario 4 the IPCC confidence is low that emissions scenario 8.5 will occur, therefore the likelihood of Scenario 4 occurring is moderately likely in the long term. However, the IPCC is confident that the emissions will fall within the 4.5 and 8.5 range.

High Moderate Low Extirpated Current 1 21 7 4 Scenario 1 0 4 23 6 Scenario 2 0 11 18 4 Scenario 3 0 4 17 12 Scenario 4 0 1 11 21 Table ES-1. Summary results from four future scenarios.

While we have data to inform us of the stressors that are likely to impact G. bartramii populations in the future, and we understand how the these stressors can impact G. bartramii, there is uncertainty regarding the exact risk of the stressors to each population because of limitations of the data, such as where and when each stressor will occur in the future and exactly which populations will be impacted. Consequently, we made the following assumptions about stressors to the mountain ranges supporting populations:

• The following sky island mountain ranges are more heavily impacted by cross border violators than other sky island ranges containing G. bartramii: Baboquivari, Chiricahua, Mule, Pajarito-Atascosa, Santa Rita, Patagonia, and Whetstone Mountains. Consequently, we assume a higher risk of wildfire to populations in these sky island mountains. • Nonnative grasses have been reported with G. bartramii in two instances, at French Joe Canyon and Juniper Flat populations. Nonnative grasses have also been reported in habitat nearby and upslope from G. bartramii in several populations. Consequently, we assume an increased likelihood of wildfire occurrence within G. bartramii populations in the Mule and Whetstone mountain ranges. • The following are sky island mountain ranges containing current and anticipated increased future recreation: the Chiricahua, Dragoon, Pajarito-Atascosa, Patagonia, Santa Rita Mountains. Consequently, we assume a higher risk of wildfire and disturbance of plants in populations within these sky islands. • Because fires have occurred recently in each of the 10 sky island mountain ranges in the United States, we assume all of sky island mountain ranges have a chance of fire from lightning ignition. When nonnative, recreation, or cross border violator presence is also a risk in a particular range, wildfire ignition probability increases.

We examined the resiliency, representation, and redundancy of G. bartramii under each of four plausible scenarios (Table ES–2). Resiliency of G. bartramii populations depends on future

vii

ground and surface water availability; availability of precipitation, especially in the winter; presence of shade to maintain microclimate, and presence of pollinators (though this factor was not analyzed). We expect the 29 extant G. bartramii populations to experience changes to these aspects of their habitat in different ways under the different scenarios. We projected the likelihood of each scenario based on the events that would occur under each scenario (Table ES– 3). Table ES–4 provides our understanding of the needs of populations and the species, as well as the assessment of the current and future conditions of G. bartramii populations and the species.

viii

Scenario 3 Scenario 4 Population Current Scenario 1 – Scenario 2 – – Sky Island – Major (Canyon = Cyn) Condition Continuation Conservation Moderate Effects Effects Baboquivari Brown Cyn Moderate Low Moderate Low Low Thomas Cyn Low Low Low Low Extirpated Chiricahua Echo Cyn Moderate Low Moderate Low Low Indian Creek Extirpated Extirpated Extirpated Extirpated Extirpated Dragoon Carlink Cyn Extirpated Extirpated Extirpated Extirpated Extirpated Jordan Cyn Moderate Moderate Moderate Moderate Low Sheephead Moderate Low Low Low Low Slavin Gulch Moderate Low Low Low Extirpated Stronghold Moderate Moderate Moderate Moderate Low Cyn E. Stronghold High Moderate Moderate Moderate Moderate Cyn W. Empire Empire Mts Extirpated Extirpated Extirpated Extirpated Extirpated Mule Juniper Flat Moderate Low Low Low Extirpated Pajarito- Alamo Cyn Moderate Low Low Low Extirpated Atascosa Holden Cyn Moderate Extirpated Low Extirpated Extirpated Sycamore Cyn Moderate Moderate Moderate Moderate Low Warsaw Cyn Moderate Low Low Extirpated Extirpated Patagonia Alum Cyn Moderate Low Low Extirpated Extirpated Chimenea- Rincon Moderate Low Low Low Low Madrona Happy Valley Extirpated Extirpated Extirpated Extirpated Extirpated N. Happy Valley Moderate Low Moderate Low Low S. Santa Rita Adobe Cyn Moderate Low Low Low Extirpated Gardner Cyn Moderate Low Moderate Low Low Josephine Cyn Moderate Low Moderate Low Low Madera Cyn Moderate Low Low Extirpated Extirpated Squaw Gulch Low Extirpated Low Extirpated Extirpated Sycamore Cyn Moderate Low Low Extirpated Extirpated Temporal Moderate Low Moderate Low Extirpated Gulch Walker Cyn Moderate Low Moderate Extirpated Extirpated Whetstone Deathtrap Cyn Low Low Low Low Low French Joe Low Low Low Extirpated Extirpated Cyn Table ES–2. Graptopetalum bartramii population conditions under each scenario.

Scenario 3 – Likelihood of Current Scenario 1 – Scenario 2 – Moderate Scenario 4 – Major Scenario Condition Continuation Conservation increase in increase in effects Occurring at: effects 10 years n/a Highly likely Unlikely Somewhat likely Somewhat likely 40 years n/a Somewhat likely Somewhat likely Moderately likely Moderately likely Table ES–3. Highly likely = we are 90% sure that this scenario will occur; Moderately likely = we are 70–90 % sure that this scenario will occur; Somewhat likely = we are 50–70% sure that this scenario will occur; Unlikely= we are less than 50% sure that this scenario will occur.

Future Condition (Viability) 3Rs Needs Current Condition Projections based on future scenarios in 10 and 40 years: • High abundance populations. In each scenario all • 29 populations known • Multiple populations likely will lose to be extant over 12 subpopulations some individuals to a mountain ranges in within each variety of threats including U.S. and Mexico. population. wildfire, mining, and • 26 extant populations • Multiple groupings flooding. in U.S. within • 4 extirpated due to subpopulations. Continuation: Threats habitat drying and • Recruitment continue on current other causes in U.S. exceeding trajectory. • 3 populations in mortality. • 2 more populations are Mexico. Resiliency: • Riparian Elements extirpated (a total of 6). • 4 populations that Population (soil moisture, • 20 populations drop in have either not been (Large populations cooler condition class. revisited to attain or habitat with temperatures, Conservation current status, 3 in good condition higher humidity; • 0 more extirpated (a total Mexico and 1 in the able to withstand water within 10 of 4). U.S., presumed extant stochastic events) m). • 12 populations drop in with small numbers • Precipitation (soil condition class. (10 each) and low moisture, cooler Moderate Effects habitat condition temperatures, • 9 more populations are • Population status: higher humidity; extirpated (a total of 12). • 1 high resiliency overstory • 22 populations drop in • presence). 22 moderate condition class. resiliency • Deep shade from Major Effects • rock walls and 8 low resiliency • 16 more extirpated ( a overstory. (this includes 5 total of 20) unknown) • Substrate. • 27 populations drop in

• Sufficient condition class. pollinators. Representation: • Total of 3,756 adult Losses in individuals. Species • Genetic variation individuals. • Reduced genetic (Genetic and within and between • Genetic variation diversity across the range ecological diversity populations within and between of environmental to maintain important to populations is conditions (substrate, adaptive potential)

Future Condition (Viability) 3Rs Needs Current Condition Projections based on future scenarios in 10 and 40 years: maintain adaptive unknown. Because elevation, water source) potential. the species occurs due to extirpations of • Distribution of across 29 populations multiple populations in populations such in 12 mountain the future. that the range of ranges, we assume • Reduced genetic environmental some genetic diversity due to conditions is diversity exists contraction in size of represented. among mountain populations and loss of ranges. unique alleles. • Populations exist across the range of environmental conditions (substrate, elevation, water source). • Elevation: 1,067 to 2,042 m (3,500 to 6,700 ft). • Historically 33 populations in 13 sky islands. • Currently 29 populations in 12 sky islands. • Sky islands isolated; 14 to 42 km (8.7 to 26 mi) apart. Redundancy: • Subpopulations and • Populations separated Species groupings lost • A large number of by up to 8 km (5 mi). (Number, • Populations within sky populations • 26 populations < 150 distribution, and island mountain ranges distributed across individuals. connectivity of extirpated. the range of the • 14 populations <50 populations to • Sky island mountain species. individual. withstand ranges become more • catastrophic Subpopulations 1 to isolated due to events) 140 m (3.3 to 459 ft). extirpation • From 16 subpopulations, where area of population has been measured, area varies from 9 to 1,400m2 (0.002 to 0.35 ac); with an average of 420 m2 (0.1 ac). Table ES-4. Species Status Assessment summary for Graptopetalum bartramii.

Table of Contents

EXECUTIVE SUMMARY ...... iv CHAPTER 1. INTRODUCTION ...... 1 CHAPTER 2. INDIVIDUAL NEEDS – LIFE HISTORY AND BIOLOGY ...... 4 2.1 Taxonomy and Genetic Diversity ...... 4 2.2 Morphology ...... 4 2.3 Phenology (Seasonal Changes) and Reproduction ...... 6 2.4 Lifespan and Demographic Trends...... 8 2.5 Resource Needs (Habitat) of Individuals ...... 9 2.6 Ecology and Ecosystem ...... 11 CHAPTER 3 – POPULATION AND SPECIES NEEDS ...... 16 3.1 Historical and Current Range and Distribution ...... 16 3.2 Needs of Graptopetalum bartramii ...... 23 3.2.1 Population Resiliency ...... 23 3.2.2. Species Representation ...... 27 3.2.3 Species Redundancy...... 28 CHAPTER 4 –STRESSORS ON VIABILITY ...... 30 4.1 Mining ...... 30 4.2 Livestock, Wildlife, and Humans ...... 33 4.3 Altered Wildfire Regime ...... 37 4.4 Overutilization for Commercial, Recreational, Scientific, or Educational Purposes ...... 42 4.5 Climate Change and Drought ...... 42 Past and recent drivers of climate change ...... 43 Observed Changes in Climate ...... 44 Climate Change Assessment Methodology ...... 45 Efficacy of Model Projection ...... 46 Climate Future Projection ...... 46 4.6 Small Population Size and Lack of Connectivity ...... 57 4.7 Summary ...... 59 CHAPTER 5 – CURRENT CONDITIONS ...... 60 5.1 Introduction ...... 60

xii

5.2. Populations ...... 61 United States ...... 61 BABOQUIVARI MOUNTAINS ...... 61 CHIRICAHUA MOUNTAINS ...... 63 DRAGOON MOUNTAINS ...... 64 EMPIRE MOUNTAINS ...... 67 PAJARITO-ATASCOSA MOUNTAINS ...... 69 PATAGONIA MOUNTAINS ...... 72 RINCON MOUNTAINS ...... 73 SANTA RITA MOUNTAINS ...... 75 WHETSTONE MOUNTAINS ...... 78 Mexico ...... 79 SIERRA LAS AVIPAS ...... 79 SIERRA LA ESCUADRA ...... 79 SIERRA LA ESTANCIA ...... 79 5.3. Current Population Resiliency ...... 80 5.4 Current Species Representation ...... 86 5.5 Current Species Redundancy ...... 86 CHAPTER 6 – VIABILITY ...... 87 6.1 Introduction ...... 87 6.1.1 Scenarios Assessment ...... 88 6.2. Scenario 1 – Continuation ...... 93 6.2.1. Resiliency, Representation, and Redundancy ...... 93 6.3. Scenario 2 – Conservation ...... 96 6.3.1. Resiliency, Representation, and Redundancy ...... 96 6.4. Scenario 3 – Moderate increase in effects ...... 99 6.4.1. Resiliency, Representation, and Redundancy ...... 99 6.5 Scenario 4 – Major increase in effects ...... 102 6.5.1. Resiliency, Representation, and Redundancy ...... 102 6.6. Status Assessment Summary ...... 105 Literature Cited ...... 109 APPENDIX 1 ...... 118

xiii

APPENDIX 2 ...... 123

xiv

CHAPTER 1. INTRODUCTION

The Species Status Assessment (SSA) framework (USFWS 2015, entire) is an analytical approach to assess the species needs, current status, and future status of a species using the best available information. The SSA Framework uses the conservation biology principles of resiliency, redundancy, and representation (collectively known as the “3Rs”) as a lens to evaluate the current and future condition of the species. The result is an SSA Report that characterizes species’ ability to sustain populations in the wild over time (viability) based on the best scientific understanding of current and future abundance and distribution within the species’ ecological settings. The intent is for the SSA Report to be easily updated as new information becomes available and to support all functions of the U.S. Fish and Wildlife Service (FWS) Endangered Species Program from Candidate Assessment to Listing to Consultations to Recovery. As such, the SSA Report will be a living document upon which other documents, such as listing rules, recovery plans, and 5–year reviews, would be based if the species warrants listing under the Endangered Species Act (ESA).

Graptopetalum bartramii is a succulent plant of the Crassulaceae or stonecrop family. It is known to have historically occurred in 33 populations on 13 mountain ranges in southern Arizona and northern Mexico: 30 populations within 10 mountain ranges of southern Arizona and 1 population each within the Sierra las Avispas, Sonora, and the Sierras La Escuadra and La Estancia, Chihuahua, Mexico. The Chihuahua ranges are sometimes collectively referred to as the Sierra Madre Occidental; however, for this SSA, we use the names Sierras La Escuadra and La Estancia to differentiate these locations. We have current information indicating 3 Arizona populations have become extirpated through drying of habitat in recent years (Carlink Canyon, Empire Mountain, and Indian Creek) and a fourth population was lost for unknown reasons (Happy Valley North). We have only historical information with no count data (only presence or absence data) for 4 populations (Thomas Canyon and the 3 Mexico populations) and one subpopulation in Sycamore Canyon of the Pajarito-Atascosa Mountains. These four populations have not been revisited in recent years and are presumed to be extant but with low numbers of individuals (10 each) and habitat in Low condition.

Graptopetalum bartramii has been a candidate for listing since 1980 (45 FR 82480). The FWS received a petition in July 2010 to list G. bartramii and designate critical habitat under the ESA (Center for Biological Diversity 2010, entire). The FWS published a 90–day finding in August 2012 concluding that the petition presented substantial scientific or commercial information indicating that listing of the species may be warranted (77 FR 47352). This SSA Report for G. bartramii is intended to provide the biological support for the decision on whether or not to propose to list the species as threatened or endangered under the ESA and, if so, where to propose designating critical habitat. Importantly, the SSA Report does not result in a decision by the FWS on whether this species should be proposed for listing as a threatened or endangered species under the ESA. Instead, this SSA Report provides a review of the available information strictly related to the biological status of G. bartramii. The listing decision will be made by the FWS after reviewing this document and all relevant laws, regulations, and policies, and the results of a proposed decision will be announced in the Federal Register, with appropriate opportunities for public input.

For the purposes of this SSA, we are analyzing impacts to populations. Also, for the purpose of this SSA, we define a population as occurring within the same water course (i.e. stream) in a sky island range and within the distance pollinators can travel. Populations of G. bartramii can consist of one to several hundred individuals, though most contain fewer than one hundred plants (Ferguson 2014, entire; Ferguson 2016a, entire). A population may consist of one or more subpopulations of G. bartramii. Currently we are aware of up to 8 subpopulations in a single population; these subpopulations are separated by up to 8 kilometers (km) (5 miles [mi]). The proximity of subpopulations within a population allows for gene flow through cross-pollination and/or water movement. Within each subpopulation are groupings of plants. Groupings are separated by up to 1.7 km (1 mi).

For the purpose of this assessment, we generally define viability as the ability of G. bartramii to sustain populations in natural systems over time. Using the SSA framework (Figure 1.1), we consider what the species needs to maintain viability by characterizing the status of the species in terms of its resiliency, redundancy, and representation (Wolf et al. 2015, entire).

Figure 1.1 Species Status Assessment framework

• Resiliency describes the ability of populations to withstand stochastic events (arising from random factors). We can measure resiliency based on metrics of population health; for example, germination versus death rates and population size. Highly resilient populations are better able to withstand disturbances such as random fluctuations in germination rates (demographic stochasticity), variations in rainfall (environmental stochasticity), or the impacts of anthropogenic activities.

• Representation describes the ability of a species to adapt to changing environmental conditions. Representation can be measured by the breadth of genetic or environmental diversity within and among populations and gauges the probability that a species is capable of adapting to environmental changes. The more representation, or diversity, a species has, the more it is capable of adapting to changes (natural or human caused) in its environment. In the absence of species-specific genetic and ecological diversity information, we evaluate

representation based on the extent and variability of habitat characteristics across the geographical range.

• Redundancy describes the ability of a species to withstand catastrophic events. Measured by the number of populations, their resiliency, and their distribution (and connectivity), redundancy gauges the probability that the species has a margin of safety to withstand or can bounce back from catastrophic events (such as a rare destructive natural event or episode involving many populations; for example, wildfire or flooding).

To evaluate the biological status of G. bartramii into the future, we assessed a range of possible future conditions to allow us to consider the species’ resiliency, redundancy, and representation. This SSA Report provides a thorough assessment of biology and natural history and assesses demographic risks, stressors, and limiting factors in the context of determining the viability and risks of extinction for the species going forward.

The format for this SSA Report includes: (1) the life history, biology, and resource needs of individuals (Chapter 2); (2) The historical and current range and distribution of G. bartramii, population and species needs, and a framework for determining the distribution of resilient populations across its range for species viability (Chapter 3); (3) the likely causes of the current and future status of the species and stressors that affect the species’ viability (Chapter 4); (4) current condition of the species, including descriptions of each population (Chapter 5); and (5) a description of species viability in terms of resiliency, redundancy, and representation (Chapter 6). This document is a compilation of the best available scientific and commercial information and a description of past, present, and likely future risk factors to G. bartramii.

CHAPTER 2. INDIVIDUAL NEEDS – LIFE HISTORY AND BIOLOGY

In this chapter we provide the species’ basic biological information including taxonomic history, genetics, morphological description, and known life history traits. We then outline the resource needs of G. bartramii individuals. Here we report those aspects of the life history of G. bartramii that are important to our analysis in determining the viability of the species using resiliency, representation, and redundancy.

2.1 Taxonomy and Genetic Diversity

Graptopetalum bartramii is a succulent plant of the Crassulaceae or stonecrop family (Moran 1994, p. 192). Graptopetalum bartramii was first presumed to be a form of G. rusbyi due to similar appearance (though G. rusbyi has clustered rosettes and flowers in the spring, rather than the fall) and similar, but not overlapping, habitat. Edwin Bartram, however, felt his 1920 and 1923 collections represented a different species and sent live specimens to Dr. J. N. Rose who, in 1926, recognized G. bartramii to be distinct and named it after Mr. Bartram (Phillips et al. 1982, p. 1). Later, in 1939, T. Kearney and R. Peebles changed the name of all Arizona species in the genera Graptopetalum and Dudleya to the genus Echeveria (Kearney and Peebles 1951, pp. 358– 362; Phillips et al. 1982, p. 1). Phillips et al. (1982, p. 2) noted that most botanists recently concerned with this family separate the genera Graptopetalum and Dudleya.

Acevedo et al. (2004, entire) investigated the phylogenetic relationship of Graptopetalum and other genera of Crassulaceae. Their work clearly separates G. bartramii from other species (Acevedo et al. 2004, p.1101). The Flora of North America (2008, p. 227) recognizes Graptopetalum and Dudleya as distinct and recognizes this species as G. bartramii in the genus Graptopetalum. Based on this information as the best available scientific and commercial data, the FWS accepts this taxonomy.

Genetic analysis of G. bartramii specifically, has not been conducted. However, due to the large distance between populations, the often small numbers of individuals within populations, and the unlikelihood of pollination between populations, it is possible that genetic variability within populations may be low. However, there may be genetic diversity between populations. For this SSA we assume that there is low diversity of genetics within a population, but that different populations have some level of genetic diversity.

2.2 Morphology

Graptopetalum bartramii is a small, succulent (fleshy), acaulescent (without a stem) perennial plant (Phillips et al. 1982, p. 2; Figure 2.1). Graptopetalum bartramii has a basal rosette that is 7 to 16 centimeters (cm) (2.75 to 6.3 inches (in)) wide comprised of 20 or more flat to concave, smooth, blue-green leaves (Rose 1926, p. 2; Phillips et al. 1982, p. 2; Moran 1994, p. 192). One to seven showy inflorescences (includes stems, stalks, bracts, and flowers) up to 30.5 cm (12 in) in height are produced in equilateral panicles (pyramidal loosely branched flower cluster) (Figure 2.2). The branches of the panicles produce one to six (usually three) flowers each (Rose 1926, p. 2). Based on the size of the plant and the number of inflorescences and flowers produced, an individual plant can have up to 214 yellow-petaled with brown-to-red spotted flowers that are 2.54 cm (1.0 in) or more across (Rose 1926, p. 2; Phillips et al. 1982, p. 3; Moran 1994, p. 3;

Shohet 1999, p. 36; Ferguson 2017b, tables). The five-petaled flowers contain stamens that are reflexed (bent or turned backward) after pollen is shed (Rose 1926, p. 2; Moran 1994, p. 192). Flower petals close following anthesis (the flowering period of a plant). The fruits are follicles (capsule that splits along one side to release seeds), with minute seeds (0.5 to 0.9 mm (0.02 to 0.04 in) in length) having little or no endosperm (tissue surrounding the embryo that provides nutrition; Shohet 1999, pp. 3, 48). Although we do not know how many seeds are produced per follicle or per plant on average, in 2016 seeds from 249 individual G. bartramii were collected for conservation. Over 50,000 seeds were collected, or roughly 200 seeds per plant, and they were reported to resemble dust. Conservation rules were applied to seed collection with no more than one third of available seeds collected.

Figure 2.1. Graptopetalum bartramii with immature inflorescence. USFWS photo September, 2014.

Figure 2.2. Graptopetalum bartramii with open inflorescence. Photo used with permission of George Ferguson, October, 2016.

2.3 Phenology (Seasonal Changes) and Reproduction

The inflorescence stalks of G. bartramii grow for 30 to 40 days, around July and August, before coming to their full height, with the flowers then opening primarily between September and November (Kearney and Peebles 1951, p. 361; Phillips et al. 1982, pp. 2, 7; Shohet 1999, p. 25). Individual flowers produce both male and female parts, but the timing of male and female flower stages differs. Individual flowers open in succession, such that the length of time each flower remains open overlaps, allowing for various stages of flowering and fruiting to be simultaneous within an individual plant for a month or more. Each individual flower remains open for approximately 5.5 days in the wild and as much as 9 days in the greenhouse (Shohet 1999, pp. 33, 35, 54) with plants going from bud to follicle within 15 days (Ferguson 2014, p. 39). Approximately the first two days are a staminate (male) stage of anthesis in which pollen is produced. The next three days are the pistillate (female) stage, in which the stamens (male part of the flower that produces pollen) are reflexed and the stigmas (female part of the flower that receives the pollen) are fully extended with the pistils (female part of the flower) receptive to the pollen (Rose, 1926, p. 2; Shohet 1999, p. 33). The two stages of floral growth may reduce the probability of self-pollination, though it likely does still occur (Ferguson, pers. comm. May 19, 2017).

The succulent rosette leaves of Graptopetalum bartramii expand or contract with fluctuations in their water content (Shohet 1999 pp. 22, 25, 36, and 39). Plants with larger rosettes produce greater numbers of flowers, and larger plants may produce more than 160 flowers. In one study, percentage of the largest mature plants blooming was 67.7 percent, compared to 19.1 percent of smaller mature plants blooming (Ferguson 2014, p. 40). Flowering is triggered by fall rains and does not occur during periods of water stress (Shohet 1999 pp. 22, 25, 36, and 39). Post- flowering, outer leaves shrivel and inner leaves become flaccid (droop) and shrink in size. Dried petals and stalks are retained after seed development, with follicles dehiscing (opening) over a period of several months from November through January (Ferguson 2014, p. 39). The old stalks can remain on the plant for sometimes more than a year, though often for only a few months (Ferguson pers. comm. June 9, 2017). Table 2.1 and Figure 2.3 denote the life cycle of G. bartramii.

February - June July - August September - November December - January Vegetative; Flowering stalk growth; Flowering (September – Flowering stalk and rosette intermittent seed seed germination November) and seed drying; seed dispersal germination pending following summer rains dispersal (November – (November – December); conditions December) intermittent seed germination pending conditions Table 2.1. Timeline of Graptopetalum bartramii life cycle.

Figure 2.3. Graptopetalum bartramii life cycle.

Graptopetalum bartramii requires pollination for reproduction. The major pollinators of G. bartramii are Sarcophaga spp. (true flies) and Musca spp. (house flies), though Apis mellifera (honey bee) may also play a role in pollination. Other species that have been noted on G. bartramii include wasps, butterflies, and Tachinidae and Bombyllidae flies (Shohet 1999, p. 41; Ferguson2014, p. 26; Ferguson 2017b, p. 13). Fertilization success is greatest in earliest opening flowers, possibly due to more pollinators being available earlier in the season, though having a long period of flowering increases overall chance of pollination (Shohet 1999, p. 57). Of the seeds produced, approximately 20 percent are viable under optimal conditions (Shohet 1999, p. 48). Because seedlings (plants less than 1.5 cm [0.6 in] in diameter) have been located in most populations, we assume pollinator availability is not a limiting factor for this species.

Given their geographic location in the landscape (i.e. in canyons with springs and streams), it is possible that seeds are transported by water and that populations may have been founded by a single individual plant or seed (Shohet 1999, p. 58). Seeds may also be dispersed via gravity and wind. Although Shohet (1999, p. 50) suggested vegetative reproduction may occur in this species, Ferguson (2015, p. 2) stated that plants rarely branch or re-root following dislodging. The Flora of North America (2008, p. 227) states the genus is not viviparous (does not reproduce from buds that form plantlets), but that G. bartramii may have solitary or rarely multiple rosettes. It is more likely rosettes are solitary, a characteristic so common that it is a feature which separates G. bartramii from G. rusbyi, a similar southern Arizona occurring species, in most floristic keys. Therefore, G. bartramii is dependent on pollinators and seed production for reproduction. Populations within and between sky island mountain ranges are very isolated, and it is unlikely that pollen is dispersed between populations or sky islands.

There is little information available regarding the seedbank of G. bartramii. In general, seed that is very tiny has evolved a requirement of sunlight for germination, as they cannot successfully emerge from deep burial (Venable and Brown 1987, p. 360). Similarly, it is thought that G. bartramii seeds reside at the soil surface beneath the litter (Shohet 1999, p. 48). It is possible that because the seed is so small, with little endosperm, mycorrhizae (the symbiotic association of a fungus with the roots of plants) may be required for seedling establishment and growth, though this has not been studied (Felger, pers. comm. June 10, 2017). Researchers at the Desert Botanical Gardens have attempted to grow G. bartramii from seed. They had no difficulty with seed germination; however, they have experienced high seedling mortality, perhaps related to a requirement for mycorrhizae for seedling establishment.

2.4 Lifespan and Demographic Trends

The lifespan of G. bartramii is thought to be approximately 5 years; this is determined from three individuals first measured in 2013 still being alive and mature in 2017 (Ferguson, 2017b, Tables 1–3; Ferguson pers. comm. Sept 27, 2017). We are aware of several populations that have been observed again at or near their originally observed location over time, indicating that these populations are persisting and are self-sustaining over a period of several decades (Ferguson 2016a, p. 28; Ferguson 2017b, p. 10). For example, in Flux Canyon, G. bartramii have been reported on 8 occasions between 1924 and 2014; in Sycamore Canyon, G. bartramii have been reported 11 times between 1950 and 2013; and in Gardner Canyon, G. bartramii individuals were found on 7 occasions between 1960 and 2015.

There are 4 populations and 1 subpopulation reported historically which have not been recently surveyed. However, the most recent surveys, sometimes decades old, indicate that the species is present, and therefore we consider these populations extant, but with low numbers (10 individuals each) and in low habitat condition. There are four populations that are now extirpated, three of these due to the drying of their habitat. There is one group of a population that we consider extirpated, likely due to dislodging from a flooding event (The Nature Conservancy 1987, p.2). There is one subpopulation that is in a different geographical location than historically reported, but is nearby. Whether the original subpopulation was extirpated and this new subpopulation established or located for the first time, or if mapping errors occurred is unclear.

Among 5 subpopulations visited in 1982 in the Pajarito-Atascosa and Patagonia Mountains in Arizona, there was moderate seedling establishment and overall low mortality reported (Phillips et al. 1982, p. 6). Among 12 subpopulations surveyed between October 2013 and July 2014 in the Baboquivari, Dragoon, Mule, Pajarito-Atascosa, Patagonia, Santa Rita, and Whetstone Mountains in Arizona, seedlings were noted in many, but not all subpopulations. A total of 131 seedlings were counted in these surveys; seedlings were typically concentrated near a mature plant (Ferguson 2014, p.6). There was 20 percent mortality among the 1,122 individuals found during these surveys; in most instances, mortality outweighed seedling production (Ferguson 2014, pp 6–39). In one instance, a population was visited in both 2014 and 2016, and seedling number fell from 53 to 3 due to desiccation (Ferguson 2017a, p. 3). Although immature and adult plant numbers fluctuate annually pending competition and other threats, seedling numbers can fluctuate even further due to the vulnerability of this life stage to desiccation, coupled with

competition and other threats. Because of this, we focus our analysis of population size on the number of adult plants in each population, rather than total plant number.

2.5 Resource Needs (Habitat) of Individuals

Graptopetalum bartramii have been found within 10 sky island mountain ranges of southern Arizona and 3 mountain ranges in northern Mexico (SEINet). The species is typically found in deep, narrow, heavily shaded canyons with erodible soils within Madrean woodlands at elevations ranging from 1,067 to 2,042 m (3,500 to 6,700 feet (ft)). Madrean woodlands are a forested community dominated by evergreen oaks, but also containing junipers and pine trees, and characterized by mild winters and warm wet summers (Brown 1982, p. 59). Madrean evergreen woodland species include evergreen oaks with sclerophyllous (hard) leaves, such as Quercus emoryi (Emory oak), Q. arizonica (Arizona white oak), and Q. oblongifolia (Mexican blue oak). Pinus discolor (border pinyon pine) and Juniperus deppeana (alligator juniper) are also common species. Madrean evergreen woodland is typically bounded by semi-desert grasslands and savanna at warmer, drier sites in the lower-elevations, and by evergreen and broadleaf forests on more mesic and cooler sites at higher elevation, north aspect, or near riparian areas.

Graptopetalum bartramii are almost always located near water sources (springs, seeps, or intermittent streams), but above the floodline (Phillips et al. 1982, p. 4; Shohet 1999, p.22; NPS 2014, p. 2). Proximity to water may provide humidity for the plant’s microclimate. Plants are typically within 10 m (32.8 ft) from a streambed in the bottom of canyons on rocky outcrops, though can be much farther on occasion (Shohet 1999, p. 5; Ferguson 2014, p. 41; NPS 2014, p. 2; Ferguson 2016a, p. 14). The deep, narrow canyons and associated overstory species provide shade during a portion of the day, creating a cooler temperature and aiding in maintaining a humid microenvironment (Figure 2.4). In addition, the vegetation litter provides retention of soil moisture, further promoting the humid microenvironment. From a study of 12 G. bartramii subpopulations, associated species and their frequency of occurrence follow: Agave palmeri (Palmer agave; 88%), J. deppeana (81%), mosses (73%), Nolina microcarpa (sacahuista; 69%), Dasylirion wheeleri (sotol; 65%), Garrya wrightii (Wright’s silktassel; 65%), Myriopteris (=Cheilanthes) spp. (cloak fern; 58%), Q. emoryi (54%), Muhlenbergia emersleyi (bullgrass; 54%), Quercus arizonica (Arizona white oak; 42%), Pinus discolor (border pinyon; 38%), and Rhus virens (evergreen sumac; 46%; Ferguson 2014, p. 41). In addition, Cheilanthes lindheimeri (fairyswords fern) and Selaginella rupincola (rockloving spikemoss) (Figure 2.5) have been determined to be commonly associated with good G. bartramii habitat (Buckley pers. comm. December 17, 2013).

Graptopetalum bartramii root into crevices on rock ledges and cliffs on slopes of various aspect (Shohet 1999, p.22; Ferguson 2014, p. 41; NPS 2016, p. 7; Figure 2.4). Due to the location of plants in crevices or shallow soil pockets in steep canyons, adherence to substrate or soil is tenuous, and plants can be easily dislodged due to flooding, foot traffic, eroding soil, or falling rocks. Crevices provide shade, shelter, and soil moisture retention, and they offer protection from burning due to a lack of surrounding vegetation for fuel in the rocky terrain (Ferguson 2014, p. 41; Figure 2.4). Crevices above floodline offer protection from typical flood events.

Graptopetalum bartramii occur on a variety of substrates (Malusa 2001, p. 39), including volcanic (Van Devender 1981, p. 4), granitic (Phillips et al. 1982, p. 5), igneous (granite or granitoid porphyries, dacite or rhyolite, all with high content of quartz, usually as phenocrysts; Ferguson 2014, p. 41), sedimentary (sandstones and conglomerate sedimentary rocks with a high quartz content; Ferguson 2014, p. 41), and metamorphic substrates (quartzite, marble with high calcite content; Ferguson 2014, p. 41).

Based on microhabitats in which the species is typically found, species needs include crevices (with or without soil) for seeds to lodge and germinate, shade and deep leaf litter to help maintain soil moisture, and a humid microhabitat in this arid environment. The specific substrate component does not seem to be critical. In addition, for reestablishment in suitable habitat, soil seedbank may be important for this species following extended periods of drought.

Figure 2.4. Example of typical Graptopetalum bartramii habitat. USFWS photo October, 2013.

Figure 2.5. Typical Graptopetalum bartramii with associates. USFWS photo September, 2014.

2.6 Ecology and Ecosystem

Madrean evergreen woodlands of the sky islands have evolved with frequent low severity fire. Surface fires were quite common in nearly all montane forest types, including Madrean woodlands, prior to about 1900 (Swetnam et al. 2001, p. 1). The maximum interval between the relatively widespread fires typically ranged from about 10 to 30 years in the pine-dominant forests (Swetnam et al. 2001, p. 4). Due to a variety of human activities in the landscape (e.g. excessive livestock grazing, fuelwood cutting, nonnative introduction and expansion, and fire suppression starting around the turn of the last century through the mid–1900s), today these woodlands have high fuel loads, and high severity fires are becoming increasingly more common (Swetnam et al. 2001, p. 11; FireScape, 2016, entire). Swetnam et al. (2001, p. 15) note that there is no evidence that such large stand-replacing fires occurred historically; for example, fire- scar studies have revealed that only low intensity surface fire regimes persisted for the past three to five centuries. Within the 10 sky island mountain ranges known to support G. bartramii, a total of 245,873 ha (536,473 ac) have burned in wildfires over the 10 year period of 2007 to 2016 (USGS GIS data 2018a, entire), and the severity of these fires is likely greater than historical fires. In 2011 alone, over 130,000 ha (321,237 ac) of land burned between two wildfires in the Chiricahua and Pajarito-Atascosa Mountains. Table 2.2 summarizes the acres by year and illustrates a high wildfire frequency; Appendix 1 gives details on each fire, including distance to the nearest known G. bartramii population.

Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Ha 4,135 17,158 12,456 4,289 131,003 15,284 5,157 536 2,627 13,577 23,666 Ac 10,219 42,398 30,780 10,599 323,716 37,768 12,742 1,325 6,491 33,549 58,480 Table 2.2. Number of hectares (ha) and acres (ac) of land burned by wildfire within the 10 sky island mountain ranges known to support Graptopetalum bartramii during the period 2007 to 2017. Data from the United States Geological Survey 2018, entire.

As stated earlier, the Madrean woodlands have warm wet summers and mild winters. Precipitation within the sky island mountain ranges is bimodal, with winter snow and rain, and summer monsoon rain. Mean annual precipitation in the Madrean woodland habitat of southern Arizona is 250-450 mm (10-17 in), with more than 50 percent occurring in summer. The winter snow and rain coincide with G. bartramii seed germination and growth. The region has experienced a drought (a prolonged period of abnormally low rainfall) for more than 10 years (Bowers 2005, p. 421; Garfin et al. 2013, p. 3; CLIMAS 2014, entire). Winter precipitation, in particular, has been shown to have decreased over the past century, as recorded by weather stations within sky island mountain ranges containing G. bartramii (Figures 2.6a-h). Winter precipitation is needed for G. bartramii germination (though some germination likely occurs following summer rains), and both summer (July-August) and fall precipitation (captured partially in the October and November “winter” data) is needed for G. bartramii flower production. Figures 2.6 a-h depict past available precipitation data for 8 of the 10 sky island mountain ranges containing G. bartramii. Note that weather stations with longer periods of record depict overall decreasing trends (linear trend lines incorporated into figures), while shorter periods of record artificially depict increases in precipitation due to records beginning during the regional drought of the 1950s.

35.0

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1960 1970 1980 1990 2000 2010

Figure 2.6a. Average winter (October – March) precipitation for Kitt Peak, Baboquivari Mountains

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1893 1924 1934 1944 1954 1964 1974 1984 1994 2004

Figure 2.6b. Average winter (October-March) precipitation Rucker Canyon, Chiricahua Mountains.

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1893 1913 1933 1953 1973 1993 2013

Figure 2.6c. Average winter (October – March) precipitation Bisbee, Bisbee 2, Bisbee 2WNW, Mule Mountains.

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1947 1957 1967 1977 1987 1997 2007

Figure 2.6d. Average winter (October – March) Precipitation at Tumacacori National Monument, Pajarito-Atascosa Mountains.

0 1923 1943 1963 1983 2003

Figure 2.6e. Average winter (October-March) precipitation Patagonia, Patagonia Mountains.

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1992 1997 2002 2007 2012

Figure 2.6f. Average winter (October – March) precipitation Vail, sw of Rincon Mountains.

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1940 1950 1960 1970 1980 1990 2000 2010

Figure 2.6g. Average winter (October – March) precipitation Santa Rita Experimental Range, Santa Rita Mountains. Note upward trend from 1950s drought.

30.0

25.0

20.0

15.0

10.0

5.0

0.0 1910 1920 1930 1940 1950 1960 1970

Figure 2.6h. Average winter (October – March) precipitation Elgin, sw of Whetstone Mountains.

Data for graphs in Figure 2 come from the Western Region Climate Center.

CHAPTER 3 – POPULATION AND SPECIES NEEDS

In this chapter we consider the historical distribution of G. bartramii, its current distribution, and what the species needs for viability. We first review the historical and current information on the range, distribution, and ecology of the species. We next review the conceptual needs of the species, including population resiliency, redundancy, and representation to support viability and reduce the likelihood of extinction.

3.1 Historical and Current Range and Distribution

We have no information regarding historical populations that would indicate a different historical range than the current range. However, there have been many changes in the southeastern Arizona landscape since the 1890s with intensive cattle grazing, water development, and fire suppression (e.g. Bahre 1991, entire). These impacts may have reduced the range or number of populations and individuals, but we have no specific information indicating this. The current range of G. bartramii includes 12 mountain ranges with 29 populations in Cochise, Pima, and Santa Cruz Counties of southern Arizona, as well as 3 mountain ranges with one population each in Sonora and Chihuahua, Mexico (Figure 3.1; Table 3.1). The populations range in elevation from 1,067 to 2,042 m (3,500 to 6,700 feet (ft)).

Figure 3.1. Sky island mountain ranges containing G. bartramii and general known locations of G. bartramii populations within southern Arizona, United States, northern Chihuahua, MX, and northern Sonora MX. The number of known individuals from each sky island mountain range is in parentheses.

Populations on different sky island mountain ranges are widely separated (ranging from roughly 14 to 42 km (8.7 to 26 mi) apart). In the United States, G. bartramii populations within the 10 sky islands historically fell within a geographical area approximately 218 km by 204 km (135 by 127 miles). This geographical area extends from the furthest west to furthest east population and includes sky island mountain ranges and the intervening lowlands. The area is bounded to the north and south by populations of G. rusbyi, a similar species that is not known to overlap in occurrence with G. bartramii. There were some subpopulations lacking geographic boundaries. For these subpopulations, we extrapolated the acreage from the known subpopulations to the unknown populations. Extrapolated totals from the 16 known areas range from 0.6 to 5.3 ha, with an average of 2.0 ha (1.6 ac to 12.4 ac, with an average of 4.9 ac).

Most of the sky islands in the United States have been surveyed for G. bartramii, and it is unlikely that any large populations remain unaccounted for therein. For example, in recent surveys of 48 localities across 8 sky island mountain ranges of southern Arizona conducted by National Park Service personnel between 2014 and 2016, survey protocol included three surveyors spreading out to physically search along the canyon bottoms and slopes, wherever bedrock and other habitat indicators were observed and on routes often extending well beyond known localities of the species (Simpson et al. 2016, p. 2). Similarly, Ferguson has surveyed 19 localities across 10 sky island mountain ranges of southern Arizona between 2013 and 2016 (Ferguson 2014, entire; Ferguson 2016a, entire) and has surveyed suitable habitat in northern Mexico with no further populations located (Ferguson pers. comm. November 18, 2016). There are additional areas in Mexico that could support this species and further surveys are necessary.

The number of populations within each sky island mountain ranges from one population (e.g. Mule Mountains) to as many as eight populations (e.g. Santa Rita Mountains). Each of these populations contains from one to eight subpopulations, which can be separated by up to 8 km (5 mi). Within each subpopulation, plants grow in groups or clusters of one to eight groups, which are separated by up to 1.7 km (1 mi). Within each subpopulation, plants grow across an area of 1 to 140 m (3.3 to 459 ft) (Ferguson 2014, entire, Ferguson 2016a, p. 14). Graptopetalum bartramii typically occurs in small populations with limited numbers of individuals. Most populations contain fewer than one hundred plants (Ferguson 2014, entire; Ferguson 2016a, entire), but occasionally hundreds of plants can be found within a single population.

We are currently aware of 26 extant G. bartramii populations within the 10 mountain ranges in the United States. We assume that the three populations from three mountain ranges in Mexico are also extant. We have only historical information with no count data (only presence or absence data) from four populations, Thomas Canyon in Arizona and the three from Mexico; in addition there is one subpopulation in Sycamore Canyon that has no count data. All are presumed to be extant but with low numbers of individuals (e.g. approximately ten individuals per population) and habitat in Low condition (Table 3.2). Recent surveys of Indian Creek revealed no plants remaining and we presume this population is now extinct. We assume that the remaining four populations lacking data are low in number, because large populations likely would have been documented well in the literature; instead, they are recorded via single herbarium specimens collected sporadically over time. From the 29 extant G. bartramii populations across 12 mountain ranges, we estimate a total of 3,756 adult individuals (this number includes 10 individuals per population from the 4 populations that lack counts). The number of individuals in a given population can vary greatly from year to year and from season

to season, depending on weather and stressors present (Ferguson 2017b, pp. 8, 15). Populations with first survey dates from 2013 to 2017 are not likely new populations, but rather newly discovered, and we consider them to have occurred historically. Of the 33 historical populations, 26 (79 percent) contain fewer than 150 individuals, and 17 (52 percent) contain fewer than 50 individuals.

For this SSA, we consider the Thomas Canyon (last visited in 1982) and three Mexican populations (last visited in 1948, 1980, and 2002) to be extant, but with only a few individuals and Low habitat condition (Table 3.2). We acknowledge that our confidence in these populations being extant is low due to the dated and sparse information. The Mule Ridge subpopulation also lacks recent data, but we assume it is still present at low numbers. For these populations or subpopulations for which we are assuming only a few individuals and Low condition habitat, we recognize that this could under or overestimate our number of populations, total population size, and extent of range, but without recent data from these sites, we are unable to provide a better assessment.

The recorded Sugarloaf Mountain subpopulation location has changed, but we consider the plants still present with 21 individuals at the new locations. We consider the Empire Mountain and Carlink Canyon populations to be extirpated, as those were revisited and no plants were located. Although it is possible plants could reestablish from a seedbank, for the purposes of our analysis, we consider these populations extirpated. The southernmost grouping of plants in the Juniper Flats population we also consider extirpated, though the other groups comprising that population are extant. We recognize that there could be a seedbank at any of the extirpated sites. However, the habitat has changed in each instance such that the species is no longer supported; therefore we consider the populations extirpated. Table 3.1 identifies all populations and associated subpopulations, as well as current status and first and last survey date. Table 3.2 describes those populations or subpopulations of G. bartramii not revisited to attain current status, considered extirpated due to recent negative surveys, or for which recorded location has changed.

First Last Population Subpopulation Status Surveyed Surveyed United States Baboquivari Mountains Brown Canyon Brown Canyon Extant 2009 2016 Thomas Canyon Thomas Canyon Extant 1982 1982 Chiricahua Mountains Echo Canyon Echo Canyon Extant 2014 2016 Rhyolite Canyon Extant 2016 2016 1975 (exact 2015 (new Sugarloaf Mountain Extant location not location) found again) Indian Creek Indian Creek Canyon Extirpated 1970 2018 Dragoon Mountains Carlink Canyon Carlink Canyon Extirpated 1995 2015

First Last Population Subpopulation Status Surveyed Surveyed Jordan Canyon Jordan Canyon Extant 2014 2015 Sheepshead Sheepshead Extant 1985 2015 Slavin Gulch Lower Slavin Gulch Extant 2016 2016 Stronghold Canyon East Cochise Spring Extant 2001 2015 Park Canyon Extant 2015 2015 Stronghold Canyon West Rockfellow Dome Trail Extant 1992 2015 Stronghold Canyon West Extant 2015 2015 Stronghold Canyon – Extant 2016 2016 hanging canyon drainage Empire Mountains Empire Mountains Empire Mountains Extirpated 2007 2017 Mule Mountains Juniper Flat Juniper Flat and vicinity Extant 1997 2015 Pajarito / Atascosa Mountains Alamo Canyon Alamo Canyon Extant 1970 2016 Holden Canyon Holden Canyon Extant 2016 2016 Sycamore Canyon Montana Peak Vicinity Extant 1981 2016 Montana Canyon Extant 2016 2016 Mule Ridge Extant 1950 1967 Penasco Canyon; below Extant 2002 2016 dam Summit Motorway Extant 2001 2016 Sycamore Canyon Extant 1950 2013 Warsaw / Old Glory Warsaw Canyon Extant 1977 2016 Canyons Patagonia Mountains Alum Gulch Alum Gulch Extant 2014 2014 Flux Canyon Extant 1924 2014 Rincon Mountains Chimenea Canyon + Chimenea-Madrona Canyons Manning Camp Trail + Extant 1975 2016 Madrona Canyon Happy Valley North Happy Valley North Extirpated 2001 2016 Happy Valley South Happy Valley South Extant 2016 2016 Santa Rita Mountains Adobe Canyon Adobe Canyon Extant 2004 2015 Gardner Canyon Cave Creek Canyon Extant 1981 2015 Gardner Canyon Extant 1960 2015 Sawmill Canyon Extant 2004 2014 Josephine Canyon Bond Canyon Extant 2014 2018

First Last Population Subpopulation Status Surveyed Surveyed Josephine Canyon Extant 2011 2016 2016 and Madera Canyon Madera Canyon Extant 1973 2017 Squaw Gulch Squaw Gulch Extant 2014 2014 Sycamore Canyon Sycamore Canyon Extant 1988 2017 Temporal Gulch Temporal Gulch Extant 2016 2016 Upper Jones Canyon Extant 2016 2016 Walker Canyon Big Casa Blanca Canyon Extant 1975 1997 Walker Canyon Basin Extant 2002 2002 Whetstone Mountains Death Trap Canyon Death Trap Springs Extant 2013 2015 French Joe Canyon French Joe Canyon Extant 2016 2016 Mexico Sierra Las Avispas, Presumed Sierra Las Avispas,Sonora 2002 2002 (Nogales County) Extant Near Colonia Pacheco (in Sierra La Escuadra, Presumed the Municipio Nuevo Casas 1948 1948 Chihuahua Extant Grandes) Cuarenta Casas (NW of Las Presumed Sierra La Estancia, Chihuahua 1980 1980 Varas, Municipio Madera) Extant Table 3.1. Sky island mountain ranges with populations and subpopulations of G. bartramii; status, year first and last surveyed in the United States and Mexico. (First survey dates of 2013–2017 indicate new populations discovered as a result of recent survey efforts.)

Populations Not Recently Revisited Population Subpopulation Year Count Status Not revisited, 1982 (discovered / Thomas Canyon Thomas Canyon rare Considered last visited) extant Not revisited, Sierra Las Sierra Las Avispas, 2002 (discovered / no count made Considered Avispas,Sonora (Nogales County) last visited) extant Near Colonia Sierra La Not revisited, Pacheco (in the 1948 (discovered / Escuadra, no count made Considered Municipio Nuevo last visited) Chihuahua extant Casas Grandes) Sierra La Cuarenta Casas Not revisited, 1980 (discovered / Estancia, (NW of Las Varas, no count made Considered last visited) Chihuahua Municipio Madera) extant

Subpopulations Not Recently Revisited

Population Subpopulation Year Count Status “few plants” ; 1950 (discovered); Sycamore Canyon Mule Ridge no count made Not revisited, 1967 (last visited) Considered extant

Extirpated Population Subpopulation Year Count Status 6 individuals Carlink Canyon Carlink Canyon 1995 (discovered) found 1 individual 2001 (visited) found 2013 (visited) 0 Considered 2015 (last visited) 0 extirpated Empire Mountain Empire Mountain 2007 (discovered) no count made 2013 (visited) 2 individuals1 Considered 2017 (last visited) 01 extirpated Indian Creek Indian Creek 1970 (discovered) no count made 1971 (visited) no count made Considered 2018 (last visited) 0 extirpated Happy Valley Happy Valley 2001 (discovered) no count made North North 2015 not found Considered 2016 not found extirpated

Recorded Subpopulation Location Changed Population Subpopulation Year Count Status 1975 (exact Sugarloaf location not found Echo Canyon present Mountain again)

2014 (new 2 individuals

location) found

2015 (new 21 individuals location) Considered extant found

Table 3.2. Graptopetalum bartramii populations or subpopulations not revisited to attain current status, considered extirpated due to recent negative surveys, or location changed. 1 Moore pers. comm. May 10, 2017

It is not known how much area within the mountain range areas contains suitable habitat that may support this species. However, a recent survey of a small area of likely habitat in northern Mexico revealed no additional G. bartramii individuals (Ferguson pers. comm. November 18, 2016). Surveys are scheduled in 2017–2019 to search for G. bartramii in historical locations, as well as in suitable habitat in Mexico. In addition, suitable habitat within the Santa Catalina and Tumacacori Mountains in the United States have been surveyed recently with no G. bartramii found. It is possible the species was historically found in other areas of these mountain ranges in the United States and Mexico; however, the best scientific and commercial data available does not indicate this.

3.2 Needs of Graptopetalum bartramii

As discussed in Chapter 1, for the purpose of this assessment, we define viability as the ability of the species to sustain populations in the wild over time (in this case, 10 and 40 years). Using the SSA framework, we describe the species’ viability by characterizing the status of the species in terms of its resiliency, redundancy, and representation. In this section we analyze what the species needs in terms of population resiliency and species representation and redundancy.

3.2.1 Population Resiliency

For G. bartramii to maintain viability, its populations or some representative portion thereof must be resilient (i.e. withstand stochastic events arising from spatially and temporally random factors). Resilient G. bartramii populations must be large enough that stochastic events do not eliminate the entire population. As discussed above, we define a population of G. bartramii as one or more subpopulations that occur within the same water course, allowing for gene flow and movement through cross-pollination and/or through the movement of seeds in water. A resilient population of G. bartramii consists of multiple subpopulations, with a large number of individuals in each subpopulation, and where recruitment exceeds mortality. This allows for shared pollinators and seed dispersal between subpopulations and groups within the population, which can allow the population to recover from disturbance events and maintain or increase genetic diversity.

Environmentally stochastic events that have the potential to affect G. bartramii population resiliency include dewatering from mining activity, drought, and climate change, which can reduce humidity and proximity to water, and high severity wildfire, including resultant erosion, sedimentation, and shade removal. Habitat elements that increase resiliency to these events include riparian characteristics, precipitation, shade, and bedrock or soil pockets in rock ledges and cliffs. In addition, a number of demographic factors influence the resiliency of populations. Small population size has the potential to decrease G. bartramii population resiliency, as all stressors are exacerbated in populations with only a small number of individuals. Area of occupied habitat, abundance, number of subpopulations and recruitment thus all affect population resiliency. These demographic factors and habitat elements are discussed below and are shown in Figure 3.2.

Figure 3.2. Life history of Graptopetalum bartramii including life stages, processes, and resource needs. Population Resiliency Factors

While there are multiple demographic and environmental factors that can affect population resiliency, we focused on those factors that influence the population and for which we have sufficient data. The population resiliency factors listed below are the population-level influences we use in our assessment of the current and future condition of the populations.

Abundance – For populations of G. bartramii to be resilient, abundance should be large enough that local stochastic events do not eliminate all individuals, allowing the overall population to recover from any one event. A greater number of individuals in a population increases the chance that a portion of the population will survive. The necessary abundance or minimum viable population size is unknown; however, estimations can be attained from available data. For example, the number of adult individuals in each population is a good indicator of overall abundance. Knowing the required density of the plants needed for survival would be useful for this factor, but this information is not available. The survey data available for G. bartramii include the total number of adult, immature, and seedling plants from most locations. Although adult and immature plant numbers fluctuate annually pending competition and other threats, seedling numbers can fluctuate even further due to the vulnerability of this life stage to desiccation, coupled with competition and other threats. Because of this, we focus our analysis on the number of adult plants in each population, rather than total plant number. For this reason, abundance of G. bartramii in a given population is presented as the total number of adult individuals from all subpopulations within that population.

Number of subpopulations – For populations of G. bartramii to be resilient, they also need multiple subpopulations per population so that local stochastic events do not eliminate the entire population, allowing the population to recover from seed dispersal from other subpopulations 24

within the population. Because subpopulations can be separated by up to 8 km (5 mi), it is possible that some stressors may impact some subpopulations but not others. However, because the species requires cross pollination for increased genetic diversity, small, scattered populations are at greater risk of extirpation through lack of reproduction. The necessary number of subpopulations per population is unknown; however, estimations can be attained from available data. For example, of the 33 populations, 1 has 6 subpopulations, 3 have 3 subpopulations, 5 have 2 subpopulations, and 24 have 1 subpopulation. We therefore consider populations with 3 or more subpopulations to have high resilience, those with 2 subpopulations to have moderate resilience, and those with a single subpopulation to have low resilience.

Recruitment – Resilient G. bartramii populations must also produce and disperse seeds, establish seedlings that survive, and maintain mature reproductive individuals in the population. Current population size and abundance reflects previous influences on the population and habitat, while reproduction and recruitment reflect population trends that may be stable, increasing or decreasing in the future. For example, a large, dense population of G. bartramii that contains mostly old individuals may be able to withstand a single stochastic event, such as a drought, over the short-term, but it is not likely to remain large and dense into the future, as there are few young individuals to sustain the population over time. A population that is less dense but has many young individuals may be likely to grow denser in the future, or such a population may be lost if a single stochastic event such as a drought desiccates many seedlings at once. Because this species requires pollination, small, scattered subpopulations within a population may produce limited seeds. Consequently, recruitment may be lower than mortality and resiliency compromised. The necessary recruitment needed for a self-sustaining population is unknown. However, we assume that a population with a greater number of seedlings than dead individuals demonstrates high recruitment. Moderate recruitment is considered when the number of seedlings is approximately equal to the number of dead plants, and low recruitment is defined as the number of seedlings being fewer than the number of dead plants.

Habitat Elements that Influence Resiliency

Habitat parameters needed for resilient populations include the shady, cooler, and humid microhabitat associated with springs and drainages in deep canyons, and commonly associated with mosses and ferns; bimodal precipitation with winter snow and rain, and summer monsoon rain; shade from overstory trees and rock walls; crevices in solid bedrock or in shallow soil pockets on rock ledges and cliffs; and the presence of pollinators for cross pollination and increased genetic diversity. Habitats with appropriate levels of these parameters are considered to contribute to resiliency, while those habitats with levels outside of the appropriate ranges are considered to provide less resiliency. Habitat in Low condition is more susceptible to loss from a single stochastic event such as wildfire or drought. The presence of crevices and soil pockets, although necessary for G. bartramii survival, are presumed to occur in all populations and therefore are not calculated in the overall population resiliency, though they are somewhat captured in the riparian element category. Although adequate pollinators are an important habitat element, this element is poorly illustrated in the literature and was therefore not rated in overall population resiliency, though is somewhat captured in the reproduction category.

Riparian Elements –Resilient populations need moisture, cooler temperatures, and humidity in their microenvironment that exceeds humidity in neighboring habitat. Proximity to water (seeps,

springs, intermittant streams, or a high water table) is a factor that may contribute to population resiliency, as G. bartramii are commonly found near and above the normal floodline in the bottom of canyons on rocky outcrops, and typically within 10 m (32.8 ft) of a streambed (Shohet 1999, p. 5; Ferguson 2014, p. 41; NPS 2014, p. 2). Although the riparian element of G. bartramii habitat is often reported as intermittent streams with scattered pools, these streambeds often support vegetation considered to be riparian, such as Salix sp. (willow), Populus sp. (cottonwood), Platanus sp. (sycamore), and, on occasion, Carex sp. (sedge) (Ferguson 2014, entire; Ferguson 2016a, entire). The relationship of nearby water and G. bartramii is unknown, but it is assumed that its presence, coupled with the topography and overstory canopy that produce shade, increases humidity and lowers temperature in the fern and moss-covered microsites where the plants occur.

A lowered water table increases the distance between the water source and individual G. bartramii plants, which would impact the soil moisture and humidity of the microenvironment. While the minimum distance from water or humidity level that is needed is not known, we assume that because the majority of plants are within 10 m (32.8 ft) of a water source or riparian vegetation community (indicating water near the surface), this is a preferred habitat condition of the species. We therefore consider populations within this distance to be in High condition. Although there are some plants that are more than 20 m (65.6 ft) from a water source, this is rare. We therefore consider populations between 10 and 20 m (32.8 to 65.6 ft) from a water source or riparian vegetation community to be in Moderate condition. Populations over 30 m (65.6 ft) from a water source or riparian vegetation community are considered to be in Low condition and at greatest risk of drying out.

Precipitation – Resilient G. bartramii populations need soil moisture, cooler temperatures, and humidity in the microenvironment. One driver of these factors is precipitation, which not only directly impacts G. bartramii survival, growth, and reproduction, but also influences water availability and shade-producing vegetation in intermittent streambeds where water flows only at certain times of the year. Overstory trees may die from decreased precipitation, stress-induced increased predation, or both; ultimately altering microclimate to be more sunny and dry and no longer suitable for G. bartramii. In addition, an increase in amount of precipitation in storm events increases flooding. An increase in flood frequency or intensity could result in an increase in the number of plants dislodged. The minimum amount of precipitation needed for individual survival is unknown. However, the succulent rosette leaves of Graptopetalum bartramii expand or contract with fluctuations in their water content, and flowering does not occur during periods of water stress; rather it is triggered by summer or fall rains (Shohet 1999 pp. 22, 25, 36, and 39). Predation on succulent leaves by small mammals and insects can also increase during periods of drought. Precipitation within the sky island mountain ranges is bimodal with winter snow and rain, and summer monsoon rain. We assume that deviation from the timing and amount of precipitation would impact the resiliency of a population, because soil moisture and humidity of the microenvironment would be impacted. This would lead to increased desiccation; alterations in the timing or amount of flowering, which may result in pollinators not being present during flowering; lack of flowering; and decreased recruitment.

Precipitation was quantified using Western Region Climate Center total precipitation for the months of October-March (deemed most important for G. bartramii germination and growth) averaged over the most recent five year period in record (typically 2012–2016). To develop

High, Moderate, and Low rankings for precipitation within G. bartramii populations, 333 precipitation data points from select weather stations (weather station total precipitation by month over period of record) were examined within the range of the species. Of these, 35 points were over 12 inches of precipitation in the winter months of October to March, 148 were between 6.1 and 12 inches during these months, and the remaining 150 points were less than 6 inch of precipitation in the winter months. Using this information, we rated populations with 15.2 cm (6 in) or less of precipitation during the winter months as Low, populations with 15.5 to 30.5 cm (6.1 to 12 in) of precipitation as Moderate, and over 30.5 cm (12 in) as High.

Shade – Another component of increasing soil moisture, and maintaining cooler temperatures and a humid microenvironment, is vegetative cover. Most G. bartramii individuals are found in deep shade from rock walls and/or overstory J. deppeana, Q. emoryi, Q. arizonica, P. discolor, as well as a variety of shrubs. It is assumed that nearby shade sources retain soil moisture over a longer period of time into the dry season, increase humidity and lower temperature in the microsites where G. bartramii occurs. There are two main reports referenced to determine shade quantity in G. bartramii populations (Ferguson 2014, entire; Ferguson 2016a, entire). These reports indicate the following approximated values of cover: 30, 50, 60, 70, 80, 90, and 100 percent cover. Of these, 30, 50, and 80 percent were the most frequently noted values. Because of this, we consider populations with more than 80 percent shade to be in High condition, those with 50 to 80 percent shade to be in Moderate condition, and those with 30 to 50 percent shade to be in Low condition.

Substrate – Graptopetalum bartramii need crevices in solid bedrock or in shallow soil pockets on rock ledges and cliffs in deep, narrow canyons. Crevices are important for seed deposition and creating a microenvironment that provides shade, shelter, and soil moisture retention. They also offer protection from burning during wildfires. Because all populations contain the required substrate for the species, this element is not rated in our analysis.

Pollinators - Graptopetalum bartramii requires pollinators for pollen transfer within and between plants. Pollinators include Sarcophaga spp. (true flies) and Musca spp. (house flies), wasps, and butterflies. The European honey bee may also play a role in pollination (Shohet 1999, pp. iii, 41). Timing of male and female flower growth differs; thus self-fertilization within a single flower is unlikely. However, if pollinators are present, it is possible that flowers within a single plant may be fertilized from pollinators visiting different flowers on the same plant (Shohet 1999, p. 45). Because the species requires pollinators and is unable to reproduce vegetatively, small, scattered populations are at greater risk of extirpation through lack of reproduction. The physical clustering of numerous plants in close proximity is probably necessary for effective genetic mixing and seed production, though this has not been studied in this species. Because all populations are assumed to currently have pollinators and this information is not available in the literature, this element is not rated in our analysis.

3.2.2. Species Representation

Maintaining representation in the form of genetic or ecological diversity is important to maintain the capacity of G. bartramii to adapt to future environmental changes. There are currently known to be 29 extant G. bartramii populations within 12 mountain ranges, with 26 populations in 10 mountain ranges in the United States, and 3 mountain ranges in northern Mexico, each with

at least one small G. bartramii population. Genetic analysis of this species has not been conducted within or among populations or sky islands. However, sky island populations on different mountain ranges are widely separated (ranging from roughly 14 to 42 km (8.7 to 26 mi) apart), making cross-pollination highly unlikely, and most of the populations contain small numbers of individuals. Therefore it is possible that genetic variability within populations may be low. However, there may be genetic diversity between populations within and between the sky islands due to response to elevational and other environmental differences between locations.

Because four populations are extirpated, and we do not know the status of four populations, it is possible that there has been a loss of genetic diversity. For this SSA, we assume that there is low genetic diversity within populations but potentially greater genetic diversity between populations and sky island mountain ranges. As such, maintaining representation in the form of genetic diversity across multiple populations and sky island mountain ranges may be important to the capacity of G. bartramii to adapt to future environmental change.

Graptopetalum bartramii occupies habitats in numerous substrate types, including granite or granitoid porphyries, dacite, or rhyolite, all with high content of quartz, usually as phenocrysts, as well as sandstones and conglomerate sedimentary rocks with high quartz content, and quartzite, marble with high calcite content. The highest concentrations of the plant occur in the Dragoon and Mule Mountains at the center of the species’ range; these areas contain a mixture of substrates, but are predominantly granite (Ferguson 2014, p. 35; Ferguson 2016a, p. 10). While it does appear that G. bartramii is more typically found on hard quartz-bearing substrates of igneous or metamorphic origin, the diversity of substrates on which it can occur could be important to maintain environmental diversity. Similarly, the species is found over a relatively wide range of elevations (1,067 to 2,042 m (3,500 to 6,700 ft)), which could be important in terms of representation. Such variability in elevation could aid in survival of future environmental changes, such as warming temperatures or decreased precipitation from climate change. At a minimum, we likely need to retain populations throughout the geographic and elevational ranges of the species to maintain the overall potential genetic and environmental diversity that can maximize the species’ response to environmental changes over time.

3.2.3 Species Redundancy

Graptopetalum bartramii needs to have multiple resilient populations distributed throughout its range to provide for redundancy. The more populations, and the wider the distribution of those populations, the more redundancy the species will exhibit. Redundancy reduces the risk that a large portion of the species’ range will be negatively affected by a catastrophic natural or anthropogenic event at a given point in time. Species that are well-distributed across their historical ranges are considered less susceptible to extinction and more likely to be viable than species confined to small portions of their ranges (Carroll et al. 2010, entire). We are aware of 29 extant G. bartramii populations within 12 mountain ranges, with 26 populations in 10 mountain ranges in the United States and 3 populations in 3 mountain ranges in Mexico. Some sky island mountain ranges have several populations and others only have one; each population may have one to eight subpopulations. There is little connectivity potential between the sky island mountain ranges (separated from roughly 14 to 42 km (8.7 to 26 mi) apart); therefore, a localized threat such as dewatering from a mine or a high intensity wildfire would impact only those populations near the activity. However, regional drought could impact many populations

throughout the plant’s range. At a minimum, we likely need to retain population redundancy across multiple sky island mountain ranges throughout the species’ range to minimize impacts from catastrophic events.

CHAPTER 4 –STRESSORS ON VIABILITY

In this chapter, we evaluate the past, current, and future stressors (i.e., negative changes in the resources needed by G. bartramii) that affect the resources that G. bartramii needs for long term viability (Figure 4.1). Current and potential future stressors, along with current and future expected distribution and abundance, determine viability and, therefore, vulnerability to extinction. We organized these stressors around stressor themes and discuss the sources of those stressors.

4.1 Mining

Mining, primarily for silver, became an important part of the southwest after the region became part of the United States in 1854 (Center for Desert Archaeology 2005, pp. 131–132). By the end of the 19th century however, copper, which was in demand for electrical wire, was a focus in mining in southern Arizona. Despite the boom and bust cycle of mining throughout the region over the decades, the region remains one of the most important copper production areas in the world (Center for Desert Archaeology 2005, pp. 131–132).

Direct removal of G. bartramii individuals and substrate due to erosion, or burial of individuals, may occur due to the placement of mineral extraction sites and debris piles. This could severely impact especially small G. bartramii populations. Erosion from test pits (an excavation made to examine the subsurface conditions of a potential mine site) has been documented to remove portions of habitat occupied by G. bartramii in Flux Canyon (Phillips et al. 1982, pp. 9–10). Soil erosion can result in burying plants, eroding the soil the plant is growing in, or dislodging plants. The crevices in which the plant occurs are typically on steep slopes and are easily disturbed and damaged. Fragmentation of G. bartramii populations due to placement of mining operations and associated activities can interfere with pollination and reproduction (Rathcke and Jules 1993, p. 276). As such, fragmentation may result in the reduction of genetic vigor (e.g. reduces the resistance to pathogens, adaptation to changing environmental conditions, and ability to colonize new habitats) and seedbank availability, and may degrade habitat. This is a stressor that could impact all populations at some point in time and is considered in our analysis of the future viability of the species.

If contamination of individuals by heavy metals and/or fugitive dust (any dust created by soil disturbing activities) generated by mining operations occurs, plant growth and vigor may be reduced as a result of changes in physiological and biochemical processes (e.g. photosynthesis, respiration, transpiration, water use efficiency, leaf conductance, growth rate, vigor, and gas exchange) and reduced pollination (Phillips et al. 1982, pp. 9–10; Chibuike and Obiora 2014, p. 1; Ferguson 2014, pp.27, 42; Waser et al. 2017, p. 90).

Pollution from mining was first recognized as a potential stressor to G. bartramii in 1980 when researchers noted that pools of water at the Flux Mine smelled foul and suggested this may affect the water quality available to G. bartramii plants (Mazzoni et al. 1980, p. 2; Phillips et al. 1982, p. 10). Ferguson (2014, pp.27, 42) suggests potential contamination of silt and heavy metals downstream of tailing piles at the Flux Mine. Eddleman (2012, entire) studied heavy metals in soil and grasses in both Flux Canyon and Alum Gulch, sites known to support G. bartramii,

Figure 4.1. Graptopetalum bartramii life history, factors affecting resource needs and processes, and stressors.

among other locations in the Patagonia Mountains. She found total concentrations of cadmium, copper, lead, zinc, and other heavy metals regularly exceeded toxic levels within soils and grass roots and aerial plants parts at her study sites, including both Alum Gulch and Flux Canyon (Eddleman 2012, pp. 57, 61, 66, and 119–146). Eddleman concludes that the number of heavy metals at or above toxic levels in grasses within the Patagonia Mountains study area suggests that metals are bioavailable to plants (2012, p. 106). It is unknown if individual G. bartramii experience reduced plant growth or vigor due to pollution, heavy metals, or fugitive dust. Consequently, we are not analyzing these stressors in our assessment of future viability.

Dewatering of streams in the vicinity of mining operations may lead to overstory canopy changes and loss of shade, as well as reduction in spring and stream flow and humidity in nearby G. bartramii populations. Figure 4.2 illustrates the typical moist, shady, canyon habitat of G. bartramii that could be impacted by dewatering of streams. One such mine, the Rosemont Mine, has been proposed in the Santa Rita Mountains. The Rosemont Final Environmental Impact Statement says no G. bartramii were found in the project area or the footprint of the connected actions; however, individuals growing in the analysis area could experience indirect impacts from groundwater drawdown (USFS 2013a, p. 676). According to the Rosemont Final Environmental Impact Statement (USFS 2013a, p. 339), the proposed mine pit would create a permanent drawdown of the water table, and groundwater would flow toward the pit and be lost to evaporation. The water would be perpetually replenished in part by groundwater from the regional aquifer, and the pit would act as a hydraulic sink.

Figure 4.2 Typical moist Graptopetalum bartramii habitat that could be impacted by water withdrawal in the Santa Rita Mountains.

The G. bartramii plants growing just southwest of the proposed Rosemont Mine near the Box Canyon / Sycamore Canyon confluence is located roughly midway between the Singing Valley Road residences and Ruelas Spring, which were analyzed in the Rosemont Final Environmental Impact Statement (USFS 2013a, pp. 346–350). The predicted drawdown at the end of active mining is 0.3 to 2.1 m (0.1 to 7 ft) for the Singing Valley Road residences and 1 to 5 feet for Ruelas Spring, depending on the model used. At 20 years from the mine closure, these numbers increase to a maximum of 4.6 m and 6.1 m (15 ft and 20 ft) for the Singing Valley Road residences and Ruelas Spring, respectively. These numbers continue to increase through time to the modeled 1,000 years from the close of the mine. Given that G. bartramii is consistently found in locations with nearby springs or other water sources, the loss of groundwater at the nearby unmapped spring in Box Canyon / Sycamore Canyon confluence, between Ruelas Spring and the Singing Valley Road residences, could significantly impact these G. bartramii plants.

The habitat needs of substrate, humidity, and shade are required for the entire life cycle of the plant and can be reduced or lost due to mining activity; pollinators are required for flowering adults and can be reduced from mining activity. Direct removal or burial of individuals will kill G. bartramii of all age classes. Similarly, the reduction in shade and / or humidity will render habitat unsuitable the species. Reduction in pollinators will impact flowering adults and could cause reduction in genetic vigor (e.g. reduces the resistance to pathogens, adaptation to changing environmental conditions, and ability to colonize new habitats) and seedbank contribution. In the range of G. bartramii, there are many mining claims, trenching and exploration drilling activities, and a few active and proposed mines. Many currently undeveloped areas of locatable mineral deposits may be explored and / or mined in the future. We do not know the extent of future mine activity within the range of G. bartramii; however, a number of proposed mines are identified for development within G. bartramii habitat. The range of current and projected mining activities varies from 1 to 10 per sky island mountain range containing G. bartramii (USFS 2012, entire). The loss of water in any G. bartramii population could lead to extirpation of that population. This stressor will be assessed in our analysis of future species viability.

4.2 Livestock, Wildlife, and Humans Graptopetalum bartramii typically occur on steep slopes with erodible soils and areas susceptible to rock fall, making them particularly vulnerable to physical damage to their environment (Figure 4.3; Phillips et al. 1982, p. 10; Shohet 1999, p. 50; Ferguson 2014, p. 42; Ferguson 2016a, pp. 15, 26). While displaced plants may re-root (Shohet 1999, pp. 50–51, 60), it is more likely that these plants will not survive (Ferguson 2015, p. 2). The potential of soil disturbance and erosion within or above G. bartramii habitat or the trampling of individual G. bartramii plants may occur from a variety of activities including: livestock and wildlife movement; the placement and maintenance of infrastructure, trails, and roads; and recreationists or cross-border violators traveling along established trails or cross country (Phillips et al. 1982, p. 10; Shohet 1999, p. 60; Ferguson 2014, p. 42; NPS 2015, p. 4; Ferguson 2016a, p. 26).

Livestock The geographic location of most G. bartramii populations in steep terrain aids in protecting the plant from predation by ungulates, including livestock (USFS 2008, p. 12). However, plants may sometimes grow in situations where trampling may occur (USFS 2008, p12). Ferguson (2016a, pp. 11–12) found livestock in large congregations within Jordan Canyon (Jordan Canyon population) and commented that such congregations have the potential to trample G. bartramii.

Figure 4.3 Graptopetalum bartramii individual growing on steep slope, showing vulnerability to erosion.

No evidence was presented that trampling had occurred at this location. However, Ferguson did note that in Josephine Canyon (Josephine Canyon population), livestock were present and at least a single G. bartramii plant showed evidence of trampling (Ferguson 2016a, p. 20). Also, in Bond Canyon (Josephine Canyon population), livestock were present and several G. bartramii showed evidence of trampling (Ferguson 2016a, p. 22). Ferguson regularly reported the presence of livestock in the G. bartramii populations he visited and noted that potential trampling was a threat (Ferguson 2014, entire; Ferguson 2016a, entire).

There is speculation that historical populations growing in areas accessible to livestock may have become extirpated due to livestock congregation in wet areas and the fragile habitat in which the plant grows (Buckley pers. comm. May 9, 2017). This is demonstrated in the Flux Canyon subpopulation of the Alum Gulch population, where there were areas accessible to cattle and other areas not accessed by cattle. Where cattle accessed G. bartramii habitat, the type locality has never been relocated, possibly due to trampling at the site. In contrast, the site where the landscape prevents cattle passage had G. bartramii growing alongside the creek (Buckley pers. comm. May 9, 2017). Given the limited evidence for livestock impacts to entire populations, this stressor is considered in our analysis of future viability only when it may impact a population with fewer than 50 individuals.

Wildlife Phillips et al. (1982, p. 10) noted that a rodent burrow above one group of G. bartramii caused erosion and uprooting of plants. Rodents (e.g. rock squirrels [Spermophilus variegatus], packrats [Neotoma spp.], pocket-gophers [Thomomys bottae]), coati [Nasua narica], javelina [Pecari tajacu], and insects (e.g. walkingstick [Pseudosermyle straminea], grasshopper [Melanoplus lakinus]) have been noted consuming the leaves and flowers of this species, which could

contribute to a significant reduction in seed set when predation has been high, such as in times of drought (Phillips et al. 1982, p. 8; Shohet 1999, p. 50; Ferguson 2014, p. 42; Ferguson 2016a, pp. 22, 26; Ferguson 2017c, p. 29). Figure 4.4 a and b illustrates leaf damage from predation. Within the Alamo Canyon population of the Pajarito-Atascosa Mountains, in 2001, observers noted that several plants appeared browsed; though there was no indication what species was responsible (Heritage Database Management System, EO ID 14804.0).

b Figure 4.4 a and b. Examples of Graptopetalum bartramii leaf damage from predation.

Another insect predator to this plant is the Xami hairstreak butterfly (Xamia xami), which oviposits a single egg onto the leaves of G. bartramii. When the egg hatches, the larva burrows into the lower surface of the leaf and feeds on inner leaf tissue (Shohet 1999, p. 50). Shohet (1999, p. 50) reported noting 8 individual plants impacted by this butterfly, 4 of which died and the other 4 of which had new leaf growth from the center rosette after 6 weeks. Similarly, Ferguson (2016a, p. 26) noted seeing caterpillars of this butterfly on two plants in Jordan Canyon; these plants had drying outer leaves, though it was not known if the larvae caused the drying. Shohet (1999, p. 50) concludes that due to the small size of many of the G. bartramii

populations, a period with higher than usual numbers of predators could cause extirpations of small populations.

Increased predation by bark beetles (Coleoptera: Curculionidae) of neighboring overstory Juniperus and Pinus species in association with drought stress (Gaylord et al. 2013, p. 567) could also impact the shaded and humid microhabitat conditions needed by G. bartramii. Given that this species requires shade from overstory trees that may be lost to bark beetle predation, all populations are susceptible to this stressor. In addition, during periods of unusual bark beetle numbers, G. bartramii populations with fewer than 50 individuals may be heavily impacted by the loss of overstory trees to bark beetle attack. Because of these factors and those described above, predation by wildlife is considered within our analysis of future viability of the species.

Humans Ferguson noted that any human traffic within G. bartramii populations can cause soil erosion and plant loss, and this includes damage from researchers (Ferguson pers. comm. April 22, 2017). Shohet (1999, p. 60) found individual G. bartramii trampled in a group of plants that bordered a campsite. Westland Resources (2013, p. 19) noted that the potential placement of a trail through G. bartramii may impact individual plants. Ferguson (2016a, pp. 14–15) noted that one G. bartramii individual in a group located within 10 m (32.8 ft) of a frequently used hiking trail was covered by rock fall. The steep riparian canyons where G. bartramii occur offer recreational opportunities for the residents of southern Arizona and northern Mexico (62 FR 665, p. 683). With projected increases in drought condition and human population growth, visitation is expected to increase in the future with the desire to be near water and shade (USFS 2013b, pp. 369, 403). Recreational activities, if poorly managed, can result in reduced habitat quality by increasing soil compaction, streambank destabilization, erosion and sedimentation, the presence of invasive nonnative plant species, and recreation can result in the trampling of G. bartramii individuals (62 FR 665, p. 683). Sky island mountain ranges containing G. bartramii and with current and anticipated increased future recreation include the Chiricahua, Dragoon, Pajarito- Atascosa, Patagonia, Santa Rita Mountains (Ferguson, pers. comm. April 7, 2017; Buckley, pers. comm. April 8, 2017). Given the potential for increased human use of these areas for recreation, and those populations with fewer than 50 individuals may be heavily impacted during periods of unusual recreational use, human use of G. bartramii habitat for recreation is considered within our analysis of future viability.

The National Park Service (2014, pp. 3–4) stated that throughout its range, G. bartramii is found in the deep canyons extensively utilized by illegal border crossers, and this activity results in heavy foot traffic and localized impacts. Over the past decade or more, tens of thousands of people, known as cross-border violators, have illegally attempted crossings of the Mexico border into Arizona annually (FWS 2011, pp. 14, 18; Humane Borders, entire). With the increased presence of human activity in the backcountry, there is greater chance of trampling and erosion of individual G. bartramii and its habitat, as well as illegal campfires and increased chance of wildfires in the uplands (Duncan et al. 2010, p. 125). Figure 4.5 illustrates a surrogate for illegal border use in the region and demonstrates higher use in the western G. bartramii range. From this map and researcher notes from the field, we anticipate the following sky island mountain ranges to be more heavily impacted by cross border violators: Babocoquivari, Chiricahua, Mule, Pajarito-Atascosa, Santa Rita, Patagonia, and Whetstone Mountains. Given the current use of these areas for illegal border activity, and that populations with fewer than 50 individuals may be 36

heavily impacted by concentrations of illegal border activity, illegal border crossing is considered within our analysis of the future viability of the species.

Figure 4.5. Mortality of cross border violators (data provided by Humane Borders.org) (green dots) and known locations of Graptopetalum bartramii in southern Arizona (purple dots). Data provide a surrogate against which to assess use of sky island mountain ranges as travel corridors in southern AZ. Data represent reported mortalities between 1981 and April 30, 2017; over 99 percent of the more than 2,800 human mortalities were recorded between 2001 and 2017.

National Forest Service personnel noted that in 2013, heavy machinery removed several inches to several feet of soil, rock and vegetation from the north bank of Ruby Road just west of a group of G. bartramii and stated that this stressor is still present (Sebesta pers. comm. June 20, 2016). This threat was first considered by Phillips et al. in 1982 (p. 9), in which they stated that this population could easily be destroyed if the road were widened or blading undermined the slope where the plants were growing, causing extensive erosion. In one population, Shohet (1999, p. 60) noted erosion of the entire slope where metal piping was installed across a wash and up the slope, but the area was never restored; here over half a dozen G. bartramii individuals were lost to erosion. Because populations with fewer than 50 individual G. bartramii could be lost to erosion, scouring, and trampling from these activities, heavy machinery and infrastructure installation are considered in our analysis of future species viability.

4.3 Altered Wildfire Regime

Since the mid–1980s, wildfire frequency in western forests has nearly quadrupled compared to the average of the period 1970 to 1986 (also see section 2.6 for a discussion of the historical fire regime of G. bartramii habitat; Westerling et al. 2006, p. 941). The timing, frequency, extent, and destructiveness of wildfires are likely to continue to increase (Westerling et al. 2006, p. 943), especially given historical land management actions, an increase in fire starts from cross border violators and recreationists (e.g. campfires, cigarettes, target shooting), nonnative plant invasion, and continuing drought conditions (Westerling et al. 2006, p. 940; FireScape 2016,

entire; Fire Management Information System, 2016, p. 2; Tersey, pers. comm. March 30, 2017). Altered fire regimes can have direct and indirect impacts to G. bartramii and its habitat. Direct impacts include burning of individual G. bartramii not protected by rocky microhabitats resulting in injury, reduction in reproductive structures, or death. Indirect impact of fire on G. bartramii may include increased runoff of floodwaters, post-fire flooding, deposition of debris and sediment originating in the burned area, erosion, changes in vegetation community composition and structure, increased presence of nonnative plants, alterations in the hydrologic and nutrient cycles, and loss of overstory canopy shade essential for maintaining G. bartramii microhabitat (Griffis et al. 2000, p. 243; Crawford et al. 2001, p. 265; Hart et al. 2005, p. 167; Smithwick et al. 2005, p. 165; Stephens et al. 2014, p. 42; Ferguson 2014, p. 43; Ferguson 2016a, p. 26).

The nonnative plants in the uplands and within G. bartramii populations include species such as Eragrostis lehmanniana (Lehman’s lovegrass) and Melinis repens (rose natal), both of which have numerous advantages over native grasses. Eragrostis lehmanniana resprouts from roots and tiller nodes not killed by hot fire, is not hampered by the reduction in mycorrhizae associated with fire and erosion, is able to respond to winter precipitation when natives grasses are dormant, is able to produce copious seed earlier than native grasses, maintains larger seedbanks than native grasses, and has higher seedling survival and establishment than native grasses during periods of drought (Anable 1990, p. 49; Anable et al. 1992, p. 182; Robinett 1992, p. 101; Fernandez and Reynolds 2000, pp. 94-95; Crimmins and Comrie 2004, p. 464; Geiger and McPherson 2005, p. 896; Schussman et al. 2006, p. 589; O’Dea 2007, p. 149; Archer and Predick 2008, p.26; Mathias et al. 2013, entire). Melinis repens is capable of growing in low moisture situations, has prolific seed production, and culms that root from the nodes (Stokes et al. 2011, p. 527). Both species outcompete native plants, reduce structural and spacial diversity of habitats, and increased biomass and fuel loads, increasing the fire cycle. Nonnative grasses have been reported with G. bartramii individuals in two instances, at French Joe Canyon and Juniper Flat populations, increasing the likelihood of fire occurrence and subsequent impacts (Heritage Database Management System, EO ID 55; Simpson pers. comm. October 4, 2017).

We are aware of eleven wildfires (Alamo, Brown, Elkhorn, Hog, Horseshoe II, La Sierra, Lizard, Mule Ridge, Murphy, Soldier Basin, and Spring) that have occurred in known G. bartramii sites in the past decade, which killed some G. bartramii individuals and removed shade in some instances. Of these fires, we have post-fire G. bartramii status information on seven: the Brown, Elkhorn, Horseshoe II, Lizard, Mule Ridge, Murphy, and Soldier Basin. We do not know that status of G. bartramii populations following the Alamo, Hog, La Sierra, or Spring Fires.

1) Brown Fire The status of the Thomas Canyon G. bartramii population post–2016 Brown fire is unknown at the time of writing, however, Ferguson (2017, entire) reports no losses of individual G. bartramii to fire in the Brown Canyon population. The fire was thought to have burned at low intensity through the area and did not burn the higher overhanging branches of nearby trees, though some nearby grasses and other plants were completely consumed (Ferguson 2017a, p. 2).

2) Elkhorn Fire In the Baboquivari Range, the Elkhorn fire in May, 2009, burned in a mosaic pattern through the Brown Canyon G. bartramii population and resulted in the loss of at least one Quercus oblongifolia (Mexican blue oak), which provided shade to the microhabitat (Ferguson 2014, pp. 9–10). As this survey was nearly five years post-fire, it is unknown if individual G. bartramii were lost to the fire.

3) Horseshoe II Fire In the Chiricahua Mountains, the Horseshoe II fire of 2011 burned through the Echo Canyon population of G. bartramii, but as surveys for the plant occurred nearly five years post-fire, it is unknown if individual G. bartramii were lost to the fire. There was however, potential loss of shade from burned trees and shrubs (Ferguson 2016a, p. 13).

4) Lizard Fire In the Dragoon Mountains, the Lizard Fire of 2017 burned 6,063 ha (14,983 ac) and affected the Carlink and Jordan Canyons G. bartramii populations. New observations by Ferguson in fall 2017 indicate minimal direct fire effect on plants in Jordan Canyon, as the widespread fire reached just the outer periphery of the population where a few individuals appear to have died from fire nearby, and at least one was burned which lived. But, drought or insects may have taken a toll as other individuals died, mostly those noted stressed in fall 2016 before the fire (Ferguson 2017c, p. 32).

5) Mule Ridge Fire Researchers in 2016 found that in two portions of the Sycamore Canyon population, the Mule Ridge fire did not burn any G. bartramii individuals growing on the rock face, but did burn overstory Fraxinus velutina and Juniperus arizonica trees 2 to 10 m (6.6 to 32.8 ft) away, and some lichens on the rock cliff face suffered from heat and died (Ferguson 2016b, entire).

6) Murphy Fire Researchers in 2014 visited G. bartramii locations that burned in the 2011 Murphy Fire. As surveys for the plant occurred roughly 2.5 years post-fire, it is unknown if individual G. bartramii were lost to the fire. They noted, however, the loss of shade from burned Quercus spp. to the south of the population (Ferguson 2014, p. 15).

7) Soldier Basin Fire During G. bartramii surveys in the Santa Rita Mountain’s Flux Canyon, Ferguson (2014, pp. 28–29) noted the following impacts to G. bartramii from the 2013 Soldier Basin Fire: 12 individuals were killed by fire and 36 died of unknown causes; an additional 25 burned, but were still alive at this last assessment.

When looking at the number of acres burned per sky island mountain range in comparison to the number of adult individuals known from each sky island mountain range, it is of interest that the two largest populations occur in sky island mountain ranges that have had the fewest acres burned in the past ten years (Table 4.1). It is not known if this is a mere coincidence or of significance, as we do not have pre-fire population counts in any population to address this question. In the few instances in which a population was visited within a year of burn, one fire

was considered low severity and no G. bartramii were harmed, a second fire burned overstory shade plants, thus altering microhabitat, but no G. bartramii individuals were burned, and in the third instance, plants were both killed outright and injured.

Figure 4.6 illustrates the perimeters of wildfires between 2007 and 2017 within southern Arizona (orange) along with known populations of G. bartramii (blue). Wildfires have burned in all 10 sky island mountain ranges of southern Arizona that support G. bartramii during this time period. Fires did not burn through G. bartramii populations in all cases, but Figure 4.5 illustrates that fire could occur in any population within this 10-year timeframe. In addition, Appendix 1 illustrates that 22 percent of the fires between 2007 and 2017 and within the 10 southern Arizona mountain ranges known to have historically supported G. bartramii occurred either at, or within 2 km (1.2 mi) of, known G. bartramii individuals.

The direct and indirect impacts of wildfire on G. bartramii could potentially cause extirpation of small populations throughout the range of the species and have negative impacts on larger populations. Given the potential for impacts to the species from unnatural high severity wildfire in all populations, this stressor is considered in our analysis of future viability of the species.

Mountain Range Hectares Acres # Adults Pajarito-Atascosa 136,496 55,238 452 Baboquivari 58,246 23,571 122 Chiricahua 278,925 112,877 186 Dragoon 15,190 6,147 1190 Empire 2,746 1,111 0 Mule 125 51 798 Patagonia 31,727 12,839 175 Rincon 19,289 7,806 34 Santa Rita 12,154 4,919 547 Whetstone 2,181 883 222 Table 4.1 The number of hectares and acres burned by wildfire between 2007 and 2017 in the ten sky island mountain ranges known to support or have historically supported Graptopetalum bartramii, and the number of adult G. bartramii currently known to occur in these mountain ranges.

Figure 4.6. Wildfires of the southern Arizona sky island mountain ranges between 2007 and 2017 (orange) and known populations of Graptopetalum bartramii (blue). Data from the United States Geological Survey 2018, entire. In addition, fires in the Spring of 2018 in Dragoon and Baboquivrai are not captured.

4.4 Overutilization for Commercial, Recreational, Scientific, or Educational Purposes

Graptopetalum bartramii is an attractive and small plant that can be easily collected by gardeners and succulent enthusiasts. This stressor was first noted in 1982 when Phillips et al. (1982, pp. 4, 10, 11) did not report exact localities in their G. bartramii summary due to the possibility of illegal collection. Shohet (1999, p. 60) noted that tagged individuals were uprooted and taken from two sites in the Santa Rita Mountains, one near a campsite. Ferguson (2014, p. 42) noted that plants in close proximity to trails have higher discovery potential and are therefore more likely to be collected. In a 2016 on-line Google search for G. bartramii for sale, the author found an advertisement from a collector in Texas offering to pay cash for G. bartramii seedlings or rooted cuttings. One website notes that the similar southern Arizona occurring species, G. rusbyi, is cultivated and legally available for sale from cactus nurseries; however, G. bartramii is not (because it is more difficult to propagate and maintain in captivity) and is therefore vulnerable to collection (http://www.mineralarts .com/cactus/graptopetalum.html). As noted in the 2012 90–day finding for this species, small populations may not be able to recover from collection, especially if the mature, reproductive plants are removed. The removal of mature plants reduces the overall reproductive effort of the population, thereby reducing the overall resilience of the population. While documented instances of collection are limited, the impacts from collection can be profound for small populations. Consequently, this stressor is considered in our analysis of future viability when it impacts a population with fewer than 50 individuals.

The Xami hairstreak butterfly (Callophrys xami) is one of the rarest butterflies in Arizona, and G. bartramii is one of its food sources. In 1982, researchers reported that a portion of one G. bartramii population was collected as part of a study of this rare butterfly (Phillips et al. 1982, p. 9). Although this may have been a unique incident, the loss of G. bartramii individuals to scientific research relating to this butterfly remains a possibility. Impacts from such collection could be profound for populations with fewer than 50 individuals; therefore this stressor is considered within our analysis of future viability.

4.5 Climate Change and Drought

Climate change is an important consideration in the analysis of the future threats to G. bartramii. To appropriately consider climate change in this species assessment, it is important to understand what climate change is, the drivers of climate change, how future climate change is projected, and the efficacy of projections. The following is an overview of the IPCC’s Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC 2014, entire). It is important to understand the terminology used by IPCC to ensure clarity. Below is a description of the terminology used by IPCC, the observed climate changes and causes, and the future projected climate changes, risks, and impacts. Levels of confidence include five qualifiers: very low, low, medium, high and very high, and are typeset in italics, e.g., medium confidence.

The following terms have been used to indicate the assessed likelihood, and typeset in italics: Term Likelihood of the outcome Virtually certain 99–100% probability Very likely 90–100% probability Likely 66–100% probability About as likely as not 33–66% probability Unlikely 0–33% probability Very unlikely 0–10% probability Exceptionally unlikely 0–1% probability Table 4.2. Definition of likelihood used by IPCC 2014 entire. * Additional terms (extremely likely: 95–100% probability, more likely than not: >50–100% probability, and extremely unlikely:0–5% probability) may also be used when appropriate.

The degree of certainty in each key finding of the assessment is based on the type, amount, quality and consistency of evidence (e.g., data, mechanistic understanding, theory, models, expert judgment) and the degree of agreement. The summary terms for evidence are: limited, medium, or robust. For agreement, they are low, medium, or high.

Figure 4.7. A depiction of evidence and agreement statements and their relationship to confidence. Confidence increases toward the top right corner as suggested by the increasing strength of shading. Generally, evidence is most robust when there are multiple, consistent independent lines of high quality. (IPCC 2014 entire)

Past and recent drivers of climate change Natural and anthropogenic substances and processes that alter the Earth’s energy budget are drivers of climate change (IPCC 2013 p 29). Climate is affected by changes in radiative forcing due to several sources (known as radiative forcing agents). These include the concentrations of radiatively active greenhouse gases (GHGs) such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O) and chlorofluorocarbons (CFCs) (IPCC 2014 p. 44). Anthropogenic GHG emissions are mainly driven by human population size, economic activity, lifestyle, energy use, land use patterns, technology, and climate policy (IPCC 2014 p. 56). The total anthropogenic radiative forcing estimate for 2011 is substantially higher (43%) than the estimate reported in the IPCC Fourth Assessment Report (AR4) for the year 2005 (IPCC 2014 p. 44). This is caused by a combination of continued growth in most GHG concentrations and an improved estimate of radiative forcing from aerosols (IPCC 2014 p. 44). Total anthropogenic GHG emissions from 2000 to 2010 were the highest in human history and reached 49 (±4.5) GtCO2-eq/yr in 2010 (IPCC 2014 p. 45). It is extremely likely that more than half of the observed increase in global 43

average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in GHG concentrations and other anthropogenic forcings together. Cumulative emissions of CO2 largely determine global mean surface warming by the late 21st century and beyond (IPCC 2014 p.62). Most aspects of climate change will persist for many centuries even if emissions of CO2 are stopped. This represents a substantial multi-century climate change commitment created by past, present, and future emissions of CO2 (IPCC 2013 p 43).

Figure 4.8. Compatible fossil fuel emissions simulated by the CMIP5 models for four Representative Concentration Pathways (RCP) scenarios. Dashed lines represent the historical estimates and RCP emissions calculated by the Integrated Assessment Models (IAMs) used to define the RCP scenarios, and solid lines and plumes show results from CMIP5 Earth System Models (ESMs, model mean, with one standard deviation shaded).

Observed Changes in Climate Warming of the climate system is unequivocal, and since the 1950s, many of the observed changes are unprecedented over decades to millennia. The atmosphere and ocean have warmed, the amounts of snow and ice have diminished, and sea level has risen. Each of the last three decades has been successively warmer at the Earth’s surface than any preceding decade since 1850. The period from 1983 to 2012 was very likely the warmest 30–year period of the last 800 years in the Northern Hemisphere, where such assessment is possible (high confidence). The globally averaged combined land and ocean surface temperature data as calculated by a linear trend show a warming of 0.85 [0.65 to 1.06] °C 2 over the period 1880 to 2012, when multiple independently produced datasets exist. It is very likely that the number of cold days and nights has decreased and the number of warm days and nights has increased on the global scale. (IPCC 2014 p. 53)

The frequency and intensity of heavy precipitation events has likely increased in North America (IPCC 2014 p. 53). Simulated global-mean trends in the frequency of extreme warm and cold days and nights over the second half of the 20th century are generally consistent with observations (IPCC 2013 p.31). Recent detection of increasing trends in extreme precipitation and discharges in some catchments implies greater risks of flooding on a regional scale (medium confidence). Costs related to flood damage, worldwide, have been increasing since the 1970s, although this is partly due to the increasing exposure of people and assets (IPCC 2014 p. 53). Averaged over the mid-latitude land areas of the Northern Hemisphere, precipitation has increased since1901 (medium confidence before and high confidence after 1951). Impacts from recent climate-related extremes, such as heat waves, droughts, floods, cyclones and wildfires, reveal significant vulnerability and exposure of some ecosystems and many human systems to current climate variability (very high confidence) (IPCC 2014 p. 53).

Climate Change Assessment Methodology Scientists use a variety of climate models, which include consideration of natural processes and variability, as well as anthropogenic process, to evaluate observed and project future changes in climate conditions (i.e. temperature, sea level, etc.). However, there is uncertainty in these natural and anthropogenic processes and possible trajectories in the future. Consequently, a multi-model approach with a range of assumptions about the magnitude and pace of future emissions helps scientists develop different emission scenarios (Figure 4.7), upon which climate model projections are based. Together, ensembles of models, simulating the response to a range of different scenarios, map out a range of possible futures and help us understand their uncertainties.

Four emission scenarios, referred to as Representative Concentration Pathways (RCPs) were developed for the latest IPCC report (Figure 4.7, IPCC 2014 p. 57). Different emission scenarios were used in previous IPCC reports (e.g. FAR, SAR, FAR, AR4); however, new data and refined modeling has allowed for updated emission scenarios. Representative Concentration Pathways are based on a combination of integrated assessment models, simple climate models, atmospheric chemistry, and global carbon cycle models (IPCC 2014 p. 57). They are identified by their approximate total radiative forcing in year 2100 relative to 1750. These four RCPs include one stringent mitigation scenario leading to a very low forcing level (RCP2.6), two intermediate scenarios (RCP4.5 and RCP6), and one scenario with very high greenhouse gas emissions (RCP8.5). Based on the latest information, the 8.5 emissions scenario is the current trajectory (Brown and Caldeira 2017, entire).

The RCPs span the full range of radiative forcing associated with emission scenarios published in the peer-reviewed literature at the time of the development of the RCPs, and the two middle scenarios where chosen to be roughly equally spaced between the two extremes of 2.6 and 8.5. Scenarios without additional efforts to constrain emissions (“baseline scenarios”) lead to pathways ranging between RCP6 and RCP8.5 p 57. The final RCP data sets comprise land use data, harmonized GHG emissions and concentrations, gridded reactive gas and aerosol emissions, as well as ozone and aerosol abundance fields (IPCC 2014, pp. 147–148). Emission scenarios do not include natural radiative forcing such as volcanoes, only anthropogenic emissions and mitigation. Figure 4.8 demonstrates the historical emissions of carbon dioxide (CO2) and the projected emissions from the Representative Concentration Pathways (RCPs).

Figure 4.10. Four emission scenarios, referred to as Representative Concentration Pathways (RCPs) developed for the latest IPCC report (IPCC 2014, entire).

Efficacy of Model Projection The direct approach to model evaluation is to compare model output with observations and analyze the resulting difference. Even though the projections from the models were never intended to be predictions over such a short timescale, the observations through 2012 generally fall within the projections made in all past assessments (IPCC 2014, p. 131). In summary, the trend in globally averaged surface temperatures falls within the range of the previous IPCC projections. Further, there is very high confidence that models reproduce the general features of the global-scale annual mean surface temperature increase over the historical period, including the more rapid warming in the second half of the 20th century, and the cooling immediately following large volcanic eruptions.

Climate Future Projection The Southwestern United States is warming and experiencing severe droughts of extended duration, decreased stream flows, changes in amount and timing of snow melt, and changes in timing and severity of precipitation and flooding (CLIMAS 2014, entire). Southeastern Arizona and much of the American Southwest have experienced serious drought in recent decades (Bowers 2005, p. 421; Garfin et al. 2013, p. 3; CLIMAS 2014, entire), and precipitation is projected to decrease in the future with climate change, although it is expected to be more intense when it does occur (Seager et al. 2007, p. 1181; Karl et al. 2009, pp. 24, 33). Some projections suggest an overall similar amount of precipitation in the Southwest, but that it will be distributed differently in timing and intensity (Zhang et al. 2012, p. 390). Most climate change scenarios predict that the American Southwest will also become warmer during the 21st century (Overpeck et al. 2012, p. 5; Karl et al. 2009, p. 129). Graptopetalum bartramii and its habitat are very susceptible to drought, loss of humidity, increases in temperature, and increased intensity of storms and flooding (NPS 2015, p 4).

Warming Warming caused by CO2 emissions is effectively irreversible over multi-century timescales unless measures are taken to remove CO2 from the atmosphere (IPCC 2014 p.63). By the mid– 21st century the magnitudes of the projected changes are substantially affected by the choice of emissions scenario (IPCC 2014 p.59). Surface temperature is projected to rise over the 21st century under all assessed emission scenarios. Warming will continue beyond 2100 under all RCP scenarios except RCP2.6. The global mean surface temperature change for the period 46

2016–2035 relative to 1986–2005 is similar for the four RCPs, and will likely be in the range 0.3°C to 0.7°C (medium confidence) (IPCC 2014 p58) and the global mean peak surface temperature is likely 8 to 2.5 C (IPCC 2014 p 62). Relative to 1850–1900, global surface temperature change for the end of the 21st century (2081–2100) is projected to likely exceed 1.5°C for RCP4.5, RCP6 and RCP8.5 (high confidence). Warming is likely to exceed 2°C for RCP6 and RCP8.5 (high confidence), more likely than not to exceed 2°C for RCP4.5 (medium confidence), but unlikely to exceed 2°C for RCP2.6 (medium confidence) (IPCC 2014 p60). Climate system properties that determine the response to external forcing have been estimated both from climate models and from analysis of past and recent climate change. The equilibrium climate sensitivity (ECS) is likely in the range 1.5°C to 4.5°C, extremely unlikely less than 1°C, and very unlikely greater than 6°C (IPCC 2014 p.62).

Precipitation In many mid-latitude and subtropical dry regions, mean precipitation will likely decrease, while in many mid-latitude wet regions, mean precipitation will likely increase under the RCP8.5 scenario (IPCC 2014 p.60). During the last millennium, western North America drought reconstructions based on tree ring information (Figure 5.13) show longer and more severe droughts than today, particularly during the MCA in the southwestern and central United States (Meko et al., 2007) (IPCC 2013 p. 422). Projected changes in runoff (Figure 11.14c) show decreases in northern Africa, western Australia, southern Europe and southwestern USA (IPCC 2013 p.988).

Monsoon precipitation will shift later in the annual cycle; increased precipitation in extratropical cyclones will lead to large increases in wintertime precipitation over the northern third of the continent; extreme precipitation increases in tropical cyclones making landfall along the western coast of USA and Mexico, the Gulf Mexico, and the eastern coast of USA and Canada (IPCC 2013 p. 106).

Temperature Extremes It is virtually certain that there will be more frequent hot and fewer cold temperature extremes over most land areas on daily and seasonal timescales, as global mean surface temperature increases (IPCC 2014 p.58). It is very likely that heat waves will occur with a higher frequency and longer duration (IPCC 2014 p.58). Occasional cold winter extremes will continue to occur (IPCC 2014 p.60).

Precipitation Extremes It is very likely that extreme precipitation events will become more intense and frequent in many regions. Extreme precipitation events over most mid-latitude land masses and over wet tropical regions will very likely become more intense and more frequent as global mean surface temperature increases (IPCC 2014 p.60).

El Nino Over the North Pacific and North America, patterns of temperature and precipitation anomalies related to El Nino and La Nina are likely to move eastwards in the future (medium confidence). Because the saturation vapor pressure of air increases with temperature, it is expected that the amount of water vapor in air will increase with a warming climate. Globally, in all RCPs, it is likely that the area encompassed by monsoon systems will increase and monsoon precipitation is

likely to intensify and El Niño-Southern Oscillation (ENSO) related precipitation variability on regional scales will likely intensify (IPCC 2014 p.60).

Surface and Ground Water Climate change over the 21st century is projected to reduce renewable surface water and groundwater resources in most dry subtropical regions (robust evidence, high agreement), intensifying competition for water among sectors (limited evidence, medium agreement) (IPCC 2014 p.69). In presently dry regions, the frequency of droughts will likely increase by the end of the 21st century under RCP8.5 (medium confidence). In contrast, water resources are projected to increase at high latitudes (robust evidence, high agreement). The interaction of increased temperature; increased sediment, nutrient and pollutant loadings from heavy rainfall; increased concentrations of pollutants during droughts; and disruption of treatment facilities during floods will reduce raw water quality and pose risks to drinking water quality (medium evidence, high agreement) (IPCC 2014 p.69).

Winter Precipitation In most regions analyzed, it is likely that decreasing numbers of snowfall events are occurring where increased winter temperatures have been observed. Both satellite and in situ observations show significant reductions in the Northern Hemisphere snow cover extent over the past 90 years, with most of the reduction occurring in the 1980s. Because of earlier spring snowmelt, the duration of the Northern Hemisphere snow season has declined by 5.3 days per decade since the 1972/1973 winter (IPCC 2013 p. 42). Snowfall has been declining in the western USA, northeastern USA and southern margins of the seasonal snow region (IPCC 2013 p. 204).

More winter-time precipitation is falling as rain rather than snow in the western USA (Knowles et al., 2006), the Pacific Northwest and Central USA (IPCC 2013 p. 204). Many mid-latitude and subtropical arid and semi-arid regions will likely experience less precipitation. The largest precipitation changes over northern Eurasia and North America are projected to occur during the winter. (IPCC 2013 p. 44). Analyses of the southwestern USA using CMIP3 models (Seager et al., 2007) show consistent projections of drying, primarily due to a decrease in winter precipitation (IPCC 2013 p. 1080).

Soil Moisture Regional to global-scale projected decreases in soil moisture and increased agricultural drought are likely (medium confidence) in presently dry regions by the end of this century under the RCP8.5 scenario. Soil moisture drying in the Mediterranean, Southwest US, and southern African regions is consistent with projected changes in Hadley circulation and increased surface temperatures, so there is high confidence in likely surface drying in these regions by the end of this century under the RCP8.5 (IPCC 2013 p. 7).

Analysis A national Climate Explorer was developed to provide customizable graphs and maps of observed and projected temperature, precipitation, and related climate variables for every county in the contiguous United States. It is a tool to aid individuals and organizations to understand how climate conditions in their location may change over the next several decades. The program is directed and overseen by an interagency team of federal climate modeling experts, chaired by the U.S. Global Change Research Program (http://toolkit.climate.gov). Graphs in Climate

Explorer show projections generated by global climate models for the Coupled Model Intercomparison Project Phase 5 (CMIP5): projection data were statistically downscaled using the Bias-corrected Constructed Analogs method (BCCA, Hidalgo et al. 2008; Maurer et al. (2010). Under the World Climate Research Programme (WCRP) the Working Group on Coupled Modelling (WGCM) established the Coupled Model Intercomparison Project (CMIP) as a standard experimental protocol for studying the output of coupled atmosphere-ocean general circulation models (AOGCMs). CMIP provides a community-based infrastructure in support of climate model diagnosis, validation, intercomparison, documentation and data access. This framework enables a diverse community of scientists to analyze GCMs in a systematic fashion, a process which serves to facilitate model improvement. Virtually the entire international climate modeling community has participated in this project since its inception in 1995. The Program for Climate Model Diagnosis and Intercomparison (PCMDI) archives much of the CMIP data and provides other support for CMIP. PCMDI's CMIP effort is funded by the Regional and Global Climate Modeling (RGCM) Program of the Climate and Environmental Sciences Division of the U.S. Department of Energy's Office of Science, Biological and Environmental Research (BER) program.

For the maps available in the Climate Explorer, monthly data from PRISM for the historical period from 1950–2010, and high resolution climate projections were from the NASA Earth Exchange Downscaled Climate Projections at 30 arc-seconds (NEX-DCP30). These datasets are both produced at a monthly timestep and at a spatial resolution of 30 arc-seconds (800m or ~0.5 miles per pixel). Both datasets capture topographic effects on temperature and precipitation, complementing the Climate Explorer graphs with spatially rich information on future climate conditions. The NEX-DCP30 dataset was produced using the Bias-Correction Spatial Disaggregation (BCSD) statistical downscaling approach described by Thrasher et al. (2013).

Maps of past and projected mean daily maximum temperature were created using Climate Explorer for the three counties in southern Arizona in which G. bartramii occurs: Cochise, Pima, and Santa Cruz (Figure 4.10 a-c). RCP4.5 was chosen because of its intermediate GHG forcing. However, there is greater sensitivity to other forcing agents, in particular anthropogenic aerosols. We also present future projections under RCP8.5, because this is the highest emissions scenario. We did not present RCP2.6, because this scenario would not produce an effect until after the time steps we are considering in this SSA, and it is not clear that the conservation measures required to meet this emissions scenario will be implemented. The RCP4.5 emissions scenario falls between RCP4.5 and RCP8.5 and is therefore represented in this range.

Figure 4.11a. Past and projected mean daily maximum temperature in Cochise County, AZ under RCP4.5 (lower) and RCP8.5 (higher) emissions scenarios.

Figure 4.11b. Past and projected mean daily maximum temperature in Pima County, AZ under RCP4.5 (lower) and RCP8.5 (higher) emissions scenarios.

Figure 4.11c. Past and projected mean daily maximum temperature in Santa Cruz County, AZ under RCP4.5 (lower) and RCP8.5 (higher) emissions scenarios.

When temperatures rise, as has been occurring in recent decades and as is projected to continue into the future, transpiration rates also increase. Transpiration is evaporation of water from plant leaves. Higher temperatures cause stoma (the plant cells which control the openings) to open allowing water to be released to the atmosphere (United States Geological Survey 2017a, entire). Transpiration accounts for about ten percent of the moisture in the atmosphere, with the rest coming primarily from evaporation from water bodies. Succulent plants, such as G. bartramii, may have low transpiration compared with other plant species due to their thick, fleshy leaves which transpire little water. However, associated species which provide shade to G. bartramii may be impacted by higher evapotranspiration rates. Evapotranspiration rate projections under two climate scenarios (RCP 4.5 and 8.5) have been made by the United States Geological Survey and are presented for sky island mountain ranges in the west, central, and eastern portion of G. bartramii range (Figure 4.11; United States Geological Survey 2017b, entire).

Figure 4.12a. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 4.5 for the Baboquivari Mountains.

Figure 4.12b Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 8.5 for the Baboquivari Mountains.

Figure 4.12c. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 4.5 for the Santa Rita Mountains.

Figure 4.12d. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 8.5 for the Santa Rita Mountains.

Figure 4.12e. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 4.5 for the Dragoon Mountains.

Figure 4.12f. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 8.5 for the Dragoon Mountains.

Figure 4.12g. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 4.5 for the Chiricahua Mountains.

Figure 4.12h. Variability in mean monthly potential evapotranspiration under Emissions Scenario RCP 8.5 for the Chiricahua Mountains.

Continuing drought, increased temperatures, and increased evapotranspiration may lead to loss of vegetation cover and shade through the dying of overstory trees stressed from the reduction of in-stream flow (Ferguson 2014, p. 42) or from insect predation or wildfire, both of which may increase under these circumstances. Such tree mortality has already been observed in G. bartramii populations, negatively impacting available microhabitat (Ferguson 2016a, pp. 12, 17, 26).

The species is also susceptible to damage from freezing events (e.g. Austin 2012, p. 1; Ferguson 2014, pp. 23 and 40). Unusual frost events early in the season have been reported in at least one location that supports G. bartramii, Brown Canyon, in 2011 and 2013 (Cohan 2011, pers. comm. May 16, 2011; Cohan 2013, pers. comm. February 22, 2013). Frost events are not projected to decrease in severity, despite warming temperatures (Kodra et al. 2011, p. 3).

Rainfall events in the southwestern United States are projected to be less frequent but more intense, and larger flood events are expected to be more common in the future (Karl et al. 2009, p. 24). Erosion and soil loss from such storm events may increase with higher peak stream flows. Flooding can remove G. bartramii individuals occurring near the stream’s edge and has the potential to remove entire small populations (e.g. Phillips et al. 1982, p. 10; The Nature Conservancy 1987, p.2; Ferguson 2014, p. 42; Ferguson 2016a, p. 26; NPS 2015, p. 4; Ferguson 2017b, p. 15). Because continuing drought, more severe freezing events, and increased high intensity rainfall events all pose threats to G. bartramii across the range of the species, this stressor is considered in our analysis of future species viability.

Although rare species in the southwestern United States evolved with drought, recent changes in temperature, rainfall patterns, and more extreme El Niño/La Niña oscillations present stressful conditions of magnitudes greater than they faced historically and raise the question of whether

rare species can persist through these conditions (e.g., Davis and Shaw 2001, p. 673). A large fraction of species faces increased extinction risk due to climate change during and beyond the 21st century, especially as climate change interacts with other stressors (high confidence). Many climate scientists predict that numerous species will shift their geographical distributions in response to warming of the climate (McLaughlin et al. 2002, p. 6070). Most plant species cannot naturally shift their geographical ranges sufficiently fast to keep up with current and high projected rates of climate change in most landscapes will not be able to keep up at the rates projected under RCP4.5 and above in flat landscapes in this century (high confidence). Such shifts may thus result in localized extinctions over portions of the range, and, in other portions of their distributions, the occupied range may expand, depending upon habitat suitability. Changes in geographical distributions can vary from subtle to more dramatic rearrangements of occupied areas (Peterson 2003, p. 650).

Because rare species have populations that are small, fragmented, restricted to fine-scale geologic formations, and/or have limited dispersal ability, extreme climatic shifts could significantly decrease population size, increase extinction risk, and alter the distribution of the species (Davis and Shaw 2001, p. 678; Parmesan and Yohe 2003, p. 40). Additionally, populations occurring in fragmented habitats can be more vulnerable to effects of climate change and other threats, particularly for species with limited dispersal abilities (McLaughlin et al. 2002, p. 6074). Where existing occupied range is bounded by areas of unsuitable habitat, as with G. bartramii, the species’ ability to move into suitable areas is reduced, and the amount of occupied habitat could shrink accordingly. In some cases, particularly when natural movement has a high probability of failure, assisted relocation may be necessary to ensure population persistence (McLachlan et al. 2007, entire).

Many aspects of climate change and associated impacts will continue for centuries, even if anthropogenic emissions of greenhouse gases are stopped. The large fraction of anthropogenic climate change resulting from CO2 emissions is irreversible on a multi-century to millennial timescale, except in the case of a large net removal of CO2 from the atmosphere over a sustained period. Without additional mitigation efforts beyond those in place today, and even with adaptation, warming by the end of the 21st century will lead to high to very high risk of severe, widespread and irreversible impacts globally (high confidence). There are multiple mitigation pathways that are likely to limit warming to below 2°C relative to pre-industrial levels. These pathways would require substantial emissions reductions over the next few decades and near zero emissions of CO2 and other long-lived greenhouse gases by the end of the century. Implementing such reductions poses substantial technological, economic, social and institutional challenges, which increase with delays in additional mitigation and if key technologies are not available. Limiting warming to lower or higher levels involves similar challenges but on different timescales.

Emissions scenarios leading to CO2-equivalent concentrations in 2100 of about 450 ppm or lower are likely to maintain warming below 2°C over the 21st century relative to pre-industrial levels. These scenarios are characterized by 40 to 70% global anthropogenic GHG emissions reductions by 2050 compared to 2010, and emissions levels near zero or below in 2100. Mitigation scenarios reaching concentration levels of about 500 ppm CO2-eq by 2100 are more likely than not to limit temperature change to less than 2°C, unless they temporarily overshoot

concentration levels of roughly 530 ppm CO2-eq before 2100, in which case they are about as likely as not to achieve that goal.

Species Response A large fraction of the species assessed is vulnerable to extinction due to climate change, often in interaction with other threats (IPCC 2013 p. 70). Species with an intrinsically low dispersal rate, especially when occupying flat landscapes where the projected climate velocity is high, and species in isolated habitats such as mountaintops, islands or small protected areas are especially at risk (IPCC 2013 p. 70). Cascading effects through organism interactions, especially those vulnerable to phenological changes, amplify risk (high confidence) (IPCC 2013 p. 70). Recent higher air temperatures, increased evaporation, and changing precipitation patterns all negatively impact and will likely continue to impact G. bartramii, as the species is closely tied to a shady, cooler, and humid microhabitat associated with springs in deep canyons.

4.6 Small Population Size and Lack of Connectivity

Small populations are less able to recover from losses caused by random environmental changes (Shaffer and Stein 2000, pp. 308–310), such as fluctuations in reproduction (demographic stochasticity), variations in rainfall (environmental stochasticity), or changes in the frequency or severity of wildfires. Graptopetalum bartramii is known historically from 33 populations within 13 mountain ranges: 30 populations in 10 mountain ranges in southern Arizona in the United States and both the Sierra las Avispas and the Sierra Madre Occidental in northern Mexico, each with at least one population. Of the 33 known historical populations, 4 have been extirpated, 4 have no historical or current population counts (3 in Mexico and 1 in the United States). We assume that these 4 populations lacking data are low in number and estimate a number of 10 individuals per population (Figure 4.12).

The genetic diversity of the isolated populations within these mountain ranges is unknown; however, given the small population size of many populations and the distance between populations, impacts from the stressors listed in the above sections are exacerbated due to small population size. Large-scale stressors such as regional drought may increase the potential for isolation and genetic loss. In addition, because approximately 47 percent (14 populations) of the extant G. bartramii populations contain 50 or fewer individuals, loss due to erosion, trampling, collection, predation, fire, severe frost, or other stressors have the potential to seriously damage or completely remove these small populations. Although adult and immature plant numbers fluctuate annually pending competition and other threats, seedling numbers can fluctuate even further due to the vulnerability of this life stage to desiccation, coupled with competition and other threats. Among 12 subpopulations surveyed between October 2013 and July 2014 in the Baboquivari, Dragoon, Mule, Pajarito-Atascosa, Patagonia, Santa Rita, and Whetstone Mountains in Arizona, seedlings were noted in many, but not all subpopulations. A total of 131 seedlings were counted in these surveys; seedlings were typically concentrated near a mature plant (Ferguson 2014, p.6). There was 20 percent mortality among the 1,122 individuals found; in most instances, mortality outweighed seedling production (Ferguson 2014, pp 6–39). Because of this, the stressor of small population size is considered in our analysis of future species viability.

Number of Adults per Population

Juniper Flat Stronghold Cyn West Jordan Cyn Sycamore Cyn Stronghold Cyn East Echo Cyn Alum Gulch Deathtrap Cyn Alamo Cyn Josephine Cyn Sycamore Cyn Brown Cyn Gardner Cyn French Joe Cyn Adobe Cyn Madera Cyn Sheepshead Temporal Gulch Walker Cyn Happy Valley South Warsaw Cyn Thomas Cyn Sierra La Estancia Sierra La Escuadra Sierra Las Avispas Chimenea-Madrona Cyn Slavin Gulch Holden Cyn Squaw Gulch Indian Creek Happy Valley North Empire Mts Carlink Cyn 0 100 200 300 400 500 600 700 800 Figure 4.13. Total number of adult Graptopetalum bartramii known from each of the 33 current and historical United States and Mexico populations. As we have no census data for Thomas Canyon or the three locations in Mexico, we presume 10 individuals for each location. Natural data breaks at 150 and 300 individuals were used in determining population resiliency categories of Low, Moderate, and High condition. In addition, populations with fewer than 50 individuals were assessed in individual effects analysis.

The G. bartramii populations in the United States and Mexico are naturally fragmented between mountain ranges. The Baboquivari, Chiricahua, Whetstone, and Patagonia Mountains have only one or two populations and/or only have one subpopulation per population, and fewer than 200 individuals per population. Further, these sky island mountain ranges are several miles from the other sky island mountain ranges, so natural re-establishment is unlikely. In addition, the Mule Mountains contain a large numbers of individuals, but there is only one population and it is 58

approximately 38 km (23.6 mi) from the nearest population, making natural re-establishment of populations unlikely. We note also the recent discovery of G. bartramii extirpation from one of the groups of plants in this population (The Nature Conservancy 1987, p. 2; Rawoot pers. comm. September 18, 2017).

Small, reproductively isolated populations are susceptible to the loss of genetic diversity, genetic drift, and inbreeding. The loss of genetic diversity may reduce the ability of a species or population to resist pathogens and parasites, to adapt to changing environmental conditions, or to colonize new habitats. Conversely, populations that pass through a “genetic bottleneck” may subsequently benefit through the elimination of harmful alleles. Nevertheless, the net result of loss of the genetic diversity is likely to be a loss of fitness and lower chance of survival of populations and of the species. Genetic drift is a change in the frequencies of alleles in a population over time. Genetic drift can arise from random differences in founder populations and the random loss of rare alleles in small isolated populations. Genetic drift may have a neutral effect on fitness, but is also a cause of the loss of genetic diversity in small populations. Genetic drift may also result in the adaptation of an isolated population to the climates and soils of specific sites, leading to the development of distinct ecotypes and to speciation. Inbreeding depression is the loss of fitness among offspring of closely related individuals. While most animal species are susceptible to inbreeding depression, plant species vary greatly in response to inbreeding. We do not know to what extent inbreeding has reduced fitness of any G. bartramii population.

Of the 29 extant G. bartramii populations, 22 (76 percent, not including extirpated) contain fewer than 150 individuals, and 13 (45 percent, not including extirpated) contain fewer than 50 individuals each (including the 4 populations that we lack information on and assumed low number of individuals). Because impact to small populations are exacerbated by all known stressors, this is considered in our analysis of future species viability.

4.7 Summary

Our analysis of the past, current, and future stressors on the resources that G. bartramii needs for long term viability revealed that there are a number of stressors to this species (Table 4.1). The stressors that pose the largest risk to future species viability are primarily related to habitat changes: groundwater extraction from mining, long-term drought, and alteration in wildfire regime. These stressors may reduce nearby water levels, shade, and humidity within G. bartramii habitat and may directly impact individuals. Other important stressors include erosion or trampling from livestock, wildlife, or human activities; illegal collection; predation of G. bartramii or their shade trees by wildlife and insects; abnormal freezing or flooding events; or other stressors that have the potential to seriously damage or completely remove small populations. These stressors are carried forward in our assessment of the future conditions of G. bartramii populations and the viability of the species overall.

CHAPTER 5 – CURRENT CONDITIONS

5.1 Introduction

The available information indicates that there are currently 3,756 individual adult G. bartramii within 29 populations spread across 12 mountain ranges of southern Arizona and northern Mexico. We are aware of several instances where fires have swept through populations with varying degree of damage to the plants – from being unharmed to being damaged or killed. Other locations report losses of individuals from drought, flooding, predation, trampling, frost, and other factors. Most populations contain seedlings among immature and adult plants; seedlings are very susceptible to desiccation, and resurveys have shown large losses in this size class (Ferguson 2016a, p. 9; Ferguson 2017a, p 3). Table 5.1 indicates the number of populations, subpopulations, adults, and seedlings known from throughout the range of G. bartramii. Because the fate of seedlings is undetermined, our analyses are focused on the number of adult plants known.

Sub- Sky Islands Populations Populations Adults Seedlings >150 Adults populations

United States Baboquivari Mountains 2 0 122 11 1;1 Chiricahua Mountains 1 1 186 6 3;1 Dragoon Mountains 5 3 1190 574 1;1;1;1;2;3 Empire Mountains 0 0 0 0 1 Mule Mountains 1 1 798 0 1 Pajarito-Atascosa 4 1 452 61 1;1;6;1 Mountains Patagonia Mountains 1 1 175 18 2 Rincon Mountains 2 0 34 31 1;1;1 1;3;2;1;1;1; Santa Rita Mountains 8 0 547 121 2;2 Whetstone Mountains 2 0 222 3 1;1 Mexico Unknown; at Sierra Las Avispas ? 10 ? ? least 1 Near Colonia Pacheco Unknown; at (Municipio Nuevo Casas ? 10 ? ? least 1 Grandes) Cuarenta Casas (NW Las Unknown; at ? 10 ? ? Varas, Municipio Madera) least 1 At least At lease Total At least 31 At least 7 At least 46 3,756 825 Table 5.1. The number of populations, subpopulations, total adults, and total known seedlings reported from all populations of Graptopetalum bartramii in the United States and Mexico. “?” indicates information is lacking.

5.2. Populations

Unless otherwise noted, information within this section comes from the Heritage Database Management System. Most historical data comes from herbarium collection information; more recent data is largely from surveys for G. bartramii led by George Ferguson and by Steve Buckley. These surveys, like historical collections, varied in time of year the information was attained and may impact such factors presented below such as dry streambed vs. running water or pools. In addition, different survey efforts collected different data such as distance to flowing water, substrate, riparian vegetation, shade, etc. If this information is not described below for the various populations and subpopulations this is because the information is lacking. In the survey data tables in this section: ‘?’ means we do not have this information; ‘present’ means plants were found, but no population count was made, and Heritage Database Management System Element Occurrence number is given in parentheses for reference. We have limited information on the microhabitat variables at each of these G. bartramii populations, subpopulations, and groupings. What information we do have is presented below.

United States BABOQUIVARI MOUNTAINS

Figure 5.2. Baboquivari Mountains known Graptopetalum bartramii populations.

There are two G. bartramii populations in the Baboquivari Mountains: Brown Canyon (State land) and Thomas Canyon (private land) (Figure 5.2a). In total, the Baboquivari Mountains are known to support 112 adult and 11 seedlings of G. bartramii plants from the Brown Canyon population (Table 5.2). No counts were ever made in Thomas Canyon and no one has reported on this population since 1982. We assume there are a small number of individuals in the Thomas Canyon population.

Population Subpopulations Year Adults Seedlings Dead Groups Notes 2009 present ? ? ? Brown Brown Canyon Canyon (HDMS-EO-58) 2014 103 55 2 3 seedling loss 2016 112 11 5 3 - desiccation? Thomas Thomas Canyon 1982 present ? ? ? "rare" Canyon (HDMS-EO-18) Table 5.2. Survey data of Baboquivari Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Brown Canyon population Graptopetalum bartramii plants in the Brown Canyon population were found in three groups, all of which experienced some level of disturbance from the 2009 Elkhorn Fire. The first group was found in soil pockets on bare rock on north facing slopes of 70 to 90 degrees (Ferguson 2014, p. 8). Plants were located between 1 and 3 m (3.3 and 9.8 ft) from a dry ravine which was 9 to 29 m (29.5 to 95 ft) from a flowing stream (Ferguson 2014, p. 8). Quercus sp. (oak) was among the overstory, which provided over 50 percent shade to the site (Ferguson 2014, p. 8). The second group of plants was found in soil pockets and cracks in bare rock on north facing slopes of 15 to 80 degrees (Ferguson 2014, p. 8). Plants were located between 2 and 8 m (63.6 and 26.2 ft) from a flowing stream. Quercus sp. was among the overstory, which provided over 50 percent shade to the site (Ferguson 2014, p. 8). Several of these plants showed damage from trampling from an unknown source. The third group of plants was found in soil pockets and cracks in bare rock on north-northwest facing slopes of 20 to 60 degrees (Ferguson 2014, p. 9). Quercus sp. was among the overstory, which provided over 50 percent shade to the site which was between 1–9 m (3.3 and 30 ft) from dry streambed (Ferguson 2014, p. 9).

In June of 2016, a second fire, the Brown Fire, burned through the Baboquivari Mountains. The fire was of low intensity and patchy at the Brown Canyon G. bartramii population, and the plants appeared to be largely unaffected, but with some scorching (Ferguson 2017a, entire).

Thomas Canyon population Graptopetalum bartramii plants in the Thomas Canyon population were found between a seep and a spring on a vertical canyon wall in a shaded Quercus woodland (Heritage Database Management System (September 2017).

CHIRICAHUA MOUNTAINS

Figure 5.2b. Chiricahua Mountains known Graptopetalum bartramii populations.

There were historically two G. bartramii populations in the Chiricahua Mountains: Echo Canyon (on National Park Service land) and Indian Creek (on National Forest Service land) (Figure 5.2b). In total, the Chiricahua Mountains are known to support 186 adult and 6 seedlings of G. bartramii plants in the Echo Canyon population (Table 5.3). No counts were ever made in Indian Creek in 1971. A recent site visit revealed no living G. bartramii at this location (Jernigan pers. comm. February 9, 2018).

Population Subpopulations Year Adults Seedlings Dead Groups Notes Echo Echo Canyon 2016 3 0 0 1 Canyon (HDMS-EO-38) Rhyolite Canyon 2016 162 6 12 2 (HDMS-EO-38) 1975 present ? ? ? nearby but Sugarloaf not in Mountain 2014 2 0 0 1 original (HDMS-EO-38) location 2015 21 0 0 1 Indian Creek 1970 present ? ? ? Indian Canyon 1971 present ? ? ? Creek extirpated (HDMS-EO-06) 2018 0 0 0 0 Table 5.3. Survey data of Chiricahua Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Echo Canyon population Graptopetalum bartramii plants in the Echo Canyon subpopulation were found in a single group spread over 1 m (3.3 ft). This group was 9 m (30 ft) from an intermittent streambed with pools. Plants were on a rock outcrop in soil pockets between the rock on a 60 degree slope with 80

percent shade from Quercus sp., Pinus sp., and canyon cliffs (Ferguson 2016a, p. 13). These plants were within 50 m of a frequently used hiking trail. The 2011 Horseshoe II Fire burned through the area, but it is unknown if individual G. bartramii were lost to the fire, as there were no pre-fire data. Researchers noted a potential loss of burned shade trees.

Graptopetalum bartramii plants in the Rhyolite Canyon subpopulation were found in soil pockets and cracks in bare rock on north-northwest facing slopes of 45 to 80 degrees (Ferguson 2016a, p. 14). Plants were located between 10 and 70 m (32.8 and 229.7 ft) from an intermittent streambed (Ferguson 2016a, p. 14). Quercus sp. and Pinus sp. were among the overstory, which provided roughly 50 percent shade to the site (Ferguson 2016a, p. 14). One G. bartramii plant was buried by rock fall at this location, which was thought to be natural from gravity moving rocks down slope.

Graptopetalum bartramii plants in the Sugarloaf Mountain subpopulation are reported in Quercus sp. and Pinus sp. woodland, near an old dam (Heritage Database Management System redacted).

Indian Creek population Habitat data were not collected in historical visits in the early 1970s. In 2018, however, the site was revisited with 5 people searching 3 hours (total of 15 hours search time) in the most likely habitats below dryer areas and along outcrops in the canyon. It was thought that a hard freeze in this area in February of 2011 may have killed all individuals in this population (Jernigan pers. comm. February 15, 2018).

DRAGOON MOUNTAINS

Figure 5.2c. Dragoon Mountains known Graptopetalum bartramii populations.

The Dragoon Mountains contain five extant G. bartramii populations, all on National Forest Service lands: Jordan Canyon, Sheepshead, Slavin Gulch, and Stronghold Canyon East and West (Figure 5.2c). In total, the Dragoon Mountains are known to support 1,190 adult and 574 seedlings of G. bartramii plants (Table 5.4). The Carlink Canyon population, which historically supported 6 individuals, has been extirpated from the Dragoon mountains from drying of the habitat.

Population Subpopulation Year Adults Seedlings Dead Groups Notes 1995 6 0 0 ? 2001 1 0 0 ? Carlink Carlink Canyon Canyon (HDMS-EO-25) 2013 0 ? ? ? not found not found; 2015 0 0 0 0 extirpated 2014 197 29 8 2 30% were Jordan Jordan Canyon reported to have outer Canyon (HDMS-EO-57) 2015 415 57 22 2 leaves drying or plants were dying “not a lot of 1985 present ? ? ? Sheephead plants” Sheephead (HDMS-EO-22) 2014 0 0 0 ? 2015 45 86 0 2 Lower Slavin Slavin Gulch 2016 9 0 0 1 Gulch (HDMS-EO-47) 2001 50 0 0 2 Stronghold 1 Cochise Spring 2013 9 0 3 2 Canyon (HDMS-EO-30) East 2015 12 0 1 2 surveyed twice 2015 131 200 1 3 in 2015 Park Canyon 2015 57 0 1 2 (HDMS-EO-42) frequent in 5x5 Stronghold Rockfellow 1992 present ? ? ? Canyon Dome Trail meter area West (HDMS-EO-24) 2015 254 130 6 3 Stronghold Canyon West 2015 76 0 2 3 (HDMS-EO-43) Stronghold Canyon – hanging canyon 2016 203 101 17 6 drainage (HDMS-EO-46) Table 5.4. History of Dragoon Mountains Graptopetalum bartramii population, subpopulation, and group occurrences. 1 Ferguson 2014 p. 33.

Carlink Canyon population Historically, the Carlink Canyon G. bartramii subpopulation was found rooted in bedrock near the Carlink Spring in cool, mesic, shaded sites and near a PVC pipe, presumably historically used to divert spring water, 1 to 3 m (3.3 to 9.8 ft) above the canyon bottom with flowing water (Heritage Database Management System; Ferguson 2014, p. 31). In 2014, however, researchers visiting the site found the creek and the spring had dried up, and no plants were located (Ferguson 2014, p. 31). In 2017, this area burned in the Lizard Fire.

Jordan Canyon population Graptopetalum bartramii plants in the Jordan Canyon subpopulation were found in two groups. The first was in soil pockets and cracks in bare rock on north-northeast facing slopes of 45 to 90 degrees (Ferguson 2016a, p. 10). Plants were located between 0.5 and 3 m (1.6 and 9.8 ft) from an intermittent streambed with pools (Ferguson 2016a, p. 10). Populus sp. and Fraxinus sp. (ash) were among the overstory which provided 50 to 90 percent shade to the site (Ferguson 2016a, p. 10). The second group of plants was found in soil pockets and cracks in bare rock on north-northeast facing slopes of 30 to 90 degrees (Ferguson 2016a, p. 12). Plants were located between 0.5 and 13 m (1.6 and 42.7 ft) from an intermittent streambed with pools. Salix sp. and Fraxinus sp. were among the overstory plants that provided 30 to 60 percent shade to the site. Both groups were affected by the 2017 Lizard Fire.

Sheepshead population Graptopetalum bartramii plants in the Sheepshead subpopulation were on steep southeast and east facing slopes in cracks in the bedrock roughly 20 m (65.6 ft) from a dry drainage. Associated plants included P. discolor and Quercus sp.

Slavin Gulch population Graptopetalum bartramii plants found in the Lower Slavin Gulch subpopulation were along a trail in the lower gulch with sandy soil in the bedrock between 5 and 15 m from the streambed. Plants were on a northwest facing slope and were partially shaded by P. discolor and Q. emoryi.

Stronghold East population Graptopetalum bartramii plants in the Cochise Spring subpopulation were found in soil pockets and cracks in bare rock on south facing slopes of 60 to 90 degrees (Ferguson 2016a, p. 7). Plants were located between 0.5 and 3 m (1.6 and 9.8 ft) from the streambed with pools of water (Ferguson 2016a, p. 7). Fraxinus sp. and other overstory species provided 50 to 80 percent shade to the site (Ferguson 2016a, p. 7). A second group of plants from the Cochise Spring subpopulation was found in soil pockets of a rock outcrop on south-southeast facing slopes of 30 to 60 degrees (Ferguson 2016a, p. 8). Here the plants were found on the edge of a much used recreation trail roughly 5 m (16.4 ft) from the intermittent stream bed; shade cover was roughly 50 percent (Ferguson 2016a, p. 8). Other researchers reported one group of plants to be over 20 m (65.6 ft) from the stream on northwest facing slopes growing on rock outcrops.

Graptopetalum bartramii plants in the Park Canyon subpopulation were in cracks of bedrock on south aspect in partial shade with water in the creek nearby. Overstory associates included P. discolor and Q. toumeyi.

Stronghold West population Graptopetalum bartramii plants in the Rockfellow Dome subpopulation were found in three groups. The first group was found in the cracks of bedrock on a north facing slope, 10 to 20 m (32.8 to 65.6 ft) from a stream with trickling water and ample shade. The second group was similarly placed, though on a south facing aspect. Associated species included Pinus sp. and Hersperocyparis arizonica (Arizona cypress). No information was provided on the third location.

Graptopetalum bartramii plants in the Stronghold West subpopulation were found on canyon walls in silt and duff near bedrock several meters from a pool of water. Plants were found along the drainage in both shade and sun with overstory components including Quercus spp.

Graptopetalum bartramii plants in the Stronghold West hanging canyon drainage subpopulation were found along a drainage of varying aspects. Plants were in silt soil beside granite bedrock, in both sun and shade.

EMPIRE MOUNTAINS

Figure 5.2d. Empire Mountains known Graptopetalum bartramii population.

The Empire Mountains G. bartramii population historically occurred on State land and consisted of a single record from 2007 in which no population count was made (Figure 5.2d). In 2013, researchers surveyed the area twice, once finding two individuals. In 2017, the population was searched for again and not found; it is now believed to have been extirpated (Table 5.5) due to drying of the habitat.

Population Subpopulation Year Adults Seedlings Dead Groups Notes 2007 present ? ? ? Empire not found Empire 2013 0 Mountains 0 0 0 surveyed twice Mountains 2013 2 (HDMS-EO-36) in 2013 not found; 2017 0 0 0 0 extirpated Table 5.5. History of the Empire Mountains Graptopetalum bartramii population. We now believe this population has been extirpated.

Empire Mountains population The Empire Mountains subpopulation was reported to be the most arid of any G. bartramii population occurring at the ecotone of the Sonoran Desert and Oak-Juniper woodland with Prosopis velutina (velvet mesquite) in the overstory (Ferguson 2014, p. 29). A pedestrian gate along the nearby fence and well-traveled road 0.25 km (0.16 mi) away allow foot traffic from hunters and recreationists. In 2017, the Mulberry Fire burned just outside the boundary of this population; the plants were searched for at that time by researchers who had visited the plants previously, and none were located.

MULE MOUNTAINS

Figure 5.2e. Mule Mountains known Graptopetalum bartramii population

There is one population of G. bartramii in the Mule Mountains: Juniper Flat (Figure 5.2e). The population occurs in the vicinity of Delgado Canyon and Juniper Flat where a few individuals were first noted in 1997. The population is found on State, Bureau of Land Management, and private lands. In total, the Mule Mountains are known to support 798 adult G. bartramii plants with no seedlings reported (Table 5.6).

Population Subpopulations Year Adults Seedlings Dead Groups Notes 1997 a few plants ? ? ? Juniper Flat Juniper Flat (HDMS-EO-26) 2014 13 1 1 2 2015 798 0 ? 7 Table 5.6. History of Mule Mountains Graptopetalum bartramii population and groups.

Juniper Flat population In 2014, Graptopetalum bartramii plants within the Juniper Flat population were found in two groups. The first group was in soil pockets and cracks in bare rock on west-southwest facing slopes of 20 to 30 degrees (Ferguson 2014, p. 35). Plants were located between 9 and 24 m (5.6 and 14.9 ft) from a dry streambed (Ferguson 2014, pp. 35). Quercus sp. and Pinus sp. provided over 30 percent shade (Ferguson 2014, p. 35). The second group of plants was found in soil pockets and cracks in bare rock on south facing slopes of 20 to 60 degrees (Ferguson 2014, p. 36). Plants were located from 8 to 9 m (26.2 to 29.5 ft) from a dry stream bed. Quercus sp., Pinus sp., and Hesperocyparis sp. (cypress) provided over 50 percent shade (Ferguson 2014, p. 36).

In 2015, more plants were located primarily on south facing slopes in bedrock with P. discolor, J. deppeana, and H. arizonica in the overstory. Notes from some of these locations indicate individuals being found from 1 to10 m (3.3 to 32.8 ft) of a stream channel, near mine tailings, near flowing stream, above a pool of water, near water seeping down slopes of canyon, and along an old road. The invasive nonnative grass, Melinis repens (Natal grass), was reported to occur within this population (Simpson pers. comm. October 4, 2017).

There was a historical grouping of G. bartramii on private lands approximately 300 m (984 ft) to the south and west of the furthest south grouping of G. bartramii in the Juniper Flat population (The Nature Conservancy 1987, p. 2). This grouping, of unknown number, was reported to have been located, but upon a second site visit, some plants were missing, presumably due to large areas of bare rock in the upper watershed and increased runoff during the monsoon season that removed the plants (The Nature Conservancy 1987, p. 2). In a site visit in September, 2017, no plants were found at any accessible location and the habitat was noted as quite warm and dry (Rawoot pers. comm. Sept 18, 2017).

PAJARITO-ATASCOSA MOUNTAINS

Figure 5.2f. Pajarito-Atascosa Mountains known Graptopetalum bartramii populations.

There are four populations of G. bartramii in the Pajarito-Atascosa Mountains, all on National Forest Service lands: Alamo Canyon, Holden Canyon, Sycamore Canyon, and Warsaw Canyon (Figure 5.2f). In total, the Pajarito-Atascosa Mountains are known to support 452 adult and 61 seedlings of G. bartramii plants (Table 5.7).

Population Subpopulations Year Adults Seedlings Dead Groups Notes 1970 present ? ? ? 1980 present ? ? ? Alamo Alamo Canyon 2001 86 0 ? ? Canyon (HDMS-EO-12) 2014 30 0 30 3 2016 134 15 3 8 Holden Holden Canyon 2016 7 1 0 2 Canyon (HDMS-EO-54) Sycamore Montana Canyon 2016 67 8 12 3 Canyon (HDMS-EO-48) "big Montana Peak 1981 present ? ? 1 Vicinity patch" (HDMS-EO-15) 2016 14 1 0 1 "few" Mule Ridge 1950 present ? ? ? plants” (HDMS-EO-37) 1967 present ? ? ? Penasco Canyon; 2002 1 0 0 1 below dam (HDMS-EO-32) 2016 38 12 1 2 2001 15 ? ? 2 Summit Motorway 2014 6 0 1 1 (HDMS-EO-28) 2016 101 22 3 2 1950 present ? ? ? scattered 1956 present ? ? ? 1972 present ? ? ? Sycamore Canyon 1973 present ? ? ? (HDMS-EO-4) 1980 12 ? ? ? 2001 6 ? ? 2 2013 78 0 1 2 Warsaw / Old Warsaw 1977 present ? ? ? Glory Canyons Canyon ((HDMS-EO-11) 2016 13 2 0 1 Table 5.7. History of Pajarito-Atascosa Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Alamo Canyon population Graptopetalum bartramii plants in the Alamo Canyon subpopulation were found in eight groups. The first group was 2 to 8 m (6.6 to 26.2 ft) from a dry streambed in soil on rock cracks and in litter on north-northeast to northeast facing slopes of 70 to 90 percent (Ferguson 2014, p. 15). Plants were in 50 to 70 percent shade of overstory Quercus sp., J. deppeana, and others. A second group was found 1 to 3 m (3.3 to 9.8 ft) from a dry streambed on northeast facing slopes of 60 to 90 degrees. Both locations were burned in the Murphy Complex Fire of 2011 (Ferguson 2014, p. 15). A third group was found 0.5 to 4 m (1.6 to 13.1 ft) from flowing stream on north- northeast facing slopes of 45 degrees (Ferguson 2014, p. 16). These plants were over 50 percent shaded by Quercus sp. and canyon walls and growing in soil pockets of gravel talus (Ferguson

2014, p. 16). We have little information on the remaining groups, save that plants were growing out of cracks in the bedrock with silty soil and that water was flowing in the streams and shade provided by J. deppeana and Quercus sp.

Holden Canyon population Graptopetalum bartramii plants in the Holden Canyon subpopulation were found on bedrock slopes in soil, silt, and pebbles 3 m (9.8 ft) from flowing water in a creek at the first site and 20 m (65.6 ft) from a stream at the second site. Researchers noted Carnegia gigantea (saguaro) and Quercus sp. on grassy slopes above the stream channel.

Sycamore Canyon population Graptopetalum bartramii plants in the Montana Canyon subpopulation were found growing in silt pockets in bedrock on a northeast facing slope with water flowing in the drainage. Graptopetalum bartramii plants in the Montana Peak Vicinity subpopulation were found in very moist silty soil in moist shady microclimates on north facing slopes near a drainage with running water.

Graptopetalum bartramii plants in the Mule Ridge subpopulation were found in humus and silty clay on a north facing slope among Quercus sp.

Graptopetalum bartramii plants in the Penasco Canyon subpopulation were found in a west- trending canyon on exposed bedrock along a drainage bottom with flowing water, deep pools, and Salix sp. among the overstory associates.

One group of Graptopetalum bartramii plants of the Summit Motorway subpopulation was found in a dry rock outcropping far from water or a drainage. A second group of plants was 1 m (3.3 ft) above the watercourse in pockets of silty soil and in cracks in the bedrock on a northwest facing slope. The Murphy Complex Fire burned through this subpopulation in 2011.

In 1950, Graptopetalum bartramii plants in the Sycamore Canyon subpopulation were reported to be on a north-facing slope, in silt and leaf mulch on rock cliff faces. In 1972 this subpopulation was reported to be found on rocky crevices on thin Quercus duff. In 1980 it was said to be on an east-northeast facing slope just above a stream. In 2001 it was said to be in cracks on a steep west-facing rock 2 to 4 m (6.6 to 13.1 ft) above the canyon bottom at one location and on a north-facing outcrop under a trail at a second site. Canyon walls were the primary source of shade. In 2016, plants in the Sycamore Canyon subpopulation were found in two groups. The first was found 1 to 9 m (3.3 to 29.5 ft) from a flowing stream on north- northwest facing slopes of 60 to 80 degrees (Ferguson 2014, p. 11). Plants were 50 to 90 percent shaded and growing in silt pockets in bedrock (Ferguson 2014, p. 11). Plants in the second group were found on a north facing slope of 30 to 80 degrees and with 90 percent shade from rock boulder and cliffs in three directions. Groups occurred in gravelly soil pockets and on bare rock near well-traveled trails (Ferguson 2014, p. 11).

Warsaw Canyon population Graptopetalum bartramii plants in the Warsaw/Old Glory Canyons subpopulation were found on exposed bedrock 5 to 10 m (16.4 to 32.8 ft) from a stream with Salix sp. Researchers noted C.

gigantea within sight of the population, noting this location was at the upper edge of the Lower Sonoran Desert life zone.

PATAGONIA MOUNTAINS

Figure 5.2g. Patagonia Mountains known Graptopetalum bartramii population

There is one population of G. bartramii in the Patagonia Mountains, located on National Forest Service lands: Alum Gulch, containing both Alum and Flux Canyons (Figure 5.2g). In total the Patagonia Mountains are known to support 175 adult and 18 seedlings of G. bartramii plants (Table 5.8). The isotype collection of the species was collected within Flux Canyon (Bartram, 1924; University of Arizona Herbarium and the New York Herbarium).

Population Subpopulations Year Adults Seedlings Dead Groups Notes Alum Gulch Alum Gulch 2014 52 ? ? 2 (HDMS-EO-41) 1924 present ? ? ? 1935 present ? ? ? 1954 present ? ? ? "common" 1955 present ? ? ? "common" 1980 present ? ? ? Flux Canyon 1989 present ? ? ? (HDMS-EO-5) 7 locations with 4 to 20 2013 28-140 ? ? 7 plants per location 2 additional 2014 123 18 38 3 locations mentioned Table 5.8. History of Patagonia Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Alum Gulch population Graptopetalum bartramii plants in the Alum Gulch subpopulation were found in two groups, the first on a shallow bank just above the high water mark in cracks of bedrock. The second group was 2 m (6.6 ft) upslope with plants growing on the edge of a cliff. There was water throughout the canyon, though not at the sites containing G. bartramii.

Graptopetalum bartramii plants in the Flux Canyon subpopulation were found in three groups. The first group was found 3.5 to 4.5 m (11.5 to 14.8 ft) from a flowing stream on north-northeast facing slopes of 45 to 80 degrees (Ferguson 2014, p. 27). Plants were in soil pockets and talus slope with over 50 percent shade from Quercus sp. overstory and canyon walls. The second group was 0 to 6 m (0 to 20 ft) from a dry streambed on a northeast facing slope of 10 to 30 degrees (Ferguson 2014, p. 27). Plants were in soil pockets and on bare rock with 50 percent shade from Quercus sp. and canyon walls (Ferguson2014, pp. 27–28). The first two groups were approximately very near to mine tailings reclamation (Ferguson 2014, p. 27). The third group of plants was found 0.5 to 1 m (1.6 to 3.3 ft) from a dry streambed on a north facing slope of 20 to 80 degrees (Ferguson 2014, p. 28). These plants were in soil pockets and on bare rock with over 50 percent shade from Quercus sp. and canyon walls (Ferguson2014, p. 28). These plants were downstream from the mine tailings, but many had burned in the Soldier Basin Fire of 2013.

RINCON MOUNTAINS

Figure 5.2h. Rincon Mountains known Graptopetalum bartramii populations.

There were historically three populations of G. bartramii, all on National Park Service land, in the Rincon Mountains: Chimenea-Madrona Canyons, Happy Valley North, and Happy Valley South (one subpopulation straddles National Park Service and National Forest Service lands) (Figure 5.2h). Currently, there are two populations, as Happy Valley North is extirpated. In total the Rincon Mountains are known to support 34 adult and 31 seedlings of G. bartramii plants.

Chimenea-Madrona Canyons population Graptopetalum bartramii plants in the Chimenea Canyon + Manning Camp + Madrona Canyon subpopulation were found in shallow soil on bedrock outcrops, approximately 30 m from a streambed, with pools and many waterfalls in a southwest trending canyon. A large waterfall with a lot of spray was approximately 50 m (164 ft) away. Associated overstory included Quercus sp. and J. deppeana.

Happy Valley North population This population was not relocated in 2015 after 7 hours of survey (HDMS EO-35). The habitat reported was exposed granite bedrock in canyon bottoms and outcroppings at 1,372-1,676 m (4,500-5,500 ft) elevation.

Happy Valley South population At two sites, the Happy Valley South population, G. bartramii plants were found in silt soil on bedrock at stream level in partial shade from Quercus sp. and P. discolor trees (HDMS EO-49). At a third site, G. bartramii plants were found on a rock outcrop growing in soil pockets between bare rocks (HDMS EO-59). The population occurred on north to northeast facing slopes of 45 to 80 percent slope. More than 50 percent cover of shade was provided from Quercus sp., Pinus sp., Juniperus sp., and canyon slope at this location, with plants at the edge of a streambed with pools.

Population Subpopulations Year Adults Seedlings Dead Groups Notes Chimenea Canyon+Manning 1975 present ? ? ? rare Chimenea- Camp Madrona 1982 present ? ? ? Trail+Madrona Canyons 1 Canyon 2001 present ? ? ? (HDMS-EO-1) 2016 9 3 14 5 2001 present ? ? ? Happy Valley Happy Valley 2015 0 0 0 0 not found North North (HDMS-EO-35) 2016 0 0 0 0 extirpated

Happy Valley Happy South Valley Happy Valley (HDMS-EO-49 2016 25 28 11 3 Saddle; South and HDMS-EO- Miller 59) Creek Table 5.9. History of Rincon Mountains Graptopetalum bartramii population, subpopulation, and group occurrences. 1 SEINet.

SANTA RITA MOUNTAINS

Figure 5.2i. Santa Rita Mountains known Graptopetalum bartramii populations.

There are eight populations of G. bartramii in the Santa Rita Mountains; all but one occur on National Forest Service lands: Adobe Canyon (on State-owned land), Gardner Canyon, Josephine Canyon, Madera Canyon, Squaw Gulch, Sycamore Canyon, Temporal Gulch, and Walker Canyon (Figure 5.2i). In total the Santa Rita Mountains are known to support 547 adult and 121 seedlings of G. bartramii plants (Table 5.10).

Population Subpopulations Year Adults Seedlings Dead Groups Notes Adobe Adobe Canyon 2004 9 ? ? ? Canyon (HDMS-EO-45) 2015 82 99 0 3 1981 present ? ? ? Gardner Cave Creek Canyon 1997 present ? ? ? Canyon (HDMS-EO-16) 2013 0 0 0 0 not found 2015 50 0 0 4 1960 present ? ? ? 1975 present ? ? ? Gardner Canyon 1977 present ? ? ?

(HDMS-EO-7) 1982 8 ? 1 ? 2014 1 0 0 1 2015 14 0 0 ? Sawmill Canyon 2004 36 ? ? 7

(HDMS-EO-34) 2014 present ? ? ?

2014 79 2 11 1 Josephine Bond Canyon Canyon (HDMS-EO-50) 2015 98 41 28 1 2016 52 8 1 3

Josephine Canyon 2011 16 24 3 3 (HDMS-EO-51) 2016 71 12 6 2 1973 present ? ? ? 1977 present ? ? ? Madera Canyon Madera 1997 present ? ? ? (HDMS-EO-19) Canyon 2013 71 23 39 4

partial 2015 14 2 6 2 survey 2016 76 ? ? 2 Squaw Gulch Squaw Gulch 2014 5 0 0 2 (HDMS-EO-56) 1988 present ? ? ? Sycamore Sycamore Canyon 2013 22 2 86 2 Canyon (HDMS-EO-27) 2016 115 ? ? ? Temporal Temporal Gulch 2016 20 0 0 3 Gulch (HDMS-EO-52) Upper Jones Canyon 2016 7 2 1 1 (HDMS-EO-53) Big Casa Blanca 1975 6 0 0 1 Walker Canyon 1992 20 0 0 1 Canyon (HDMS-EO-23) 1997 16 0 0 1 Walker Canyon 2002 2 ? ? ? Basin (HDMS-EO- 31) 2002 3 ? ? ? Table 5.10 History of Santa Rita Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Adobe Canyon population Graptopetalum bartramii plants in the Adobe Canyon subpopulation were found in three groups. The first group was on a man-made southeast facing talus slope along the Arizona Trail. The second group was found on a northwest facing slope in soil pockets between large detached rocks just above a 15.2 m (50 ft) pour-over (waterfall area). The third group was at the confluence of a canyon and a tributary on north facing bedrock. Associates included Quercus sp. and J. deppeana in the overstory.

Gardner Canyon population Graptopetalum bartramii plants in the Cave Canyon subpopulation were found above and below a spring about 10 m (32.8 ft) from the creek bottom. Of the three sites, one was north facing and two were south facing. Associates included Platanus sp. and Quercus sp.

Graptopetalum bartramii plants in the Sawmill Canyon subpopulation were found on steep southwest facing slopes near a riparian streamside with Platanus sp. in the overstory and a spring just upstream. Plants were on cracks in bedrock and rubble at the base of a steep wall of rocks.

Graptopetalum bartramii plants in the Gardner Canyon subpopulation were found in cracks of rocks on a south facing slope along a rocky stream in Pinus – Quercus woodland.

Josephine Canyon population Graptopetalum bartramii plants in the Bond Canyon subpopulation were found in soil pockets on rock outcrop on west-northwest and north facing slopes of 45 to 90 degrees (Ferguson 2016a, p. 22). Plants were 0 to 7.5 m (0 to 24.6 ft) from an intermittent streambed with pools (Ferguson 2016a, p. 22). Shade at the site was over 50 percent from overstory including both living and dead Quercus spp. (Ferguson 2016a, p. 22). The location was near old mining road, and several hunters were encountered on the site visit (Ferguson 2016a, p. 22).

Graptopetalum bartramii plants in the Josephine Canyon subpopulation were found in two groups. The first group was on a southwest facing slope of 45 to 90 degrees, and 1 to 3 m (3.3 to 9.8 ft) from an intermittent stream bed in soil pockets and cracks of bare rock (Ferguson 2016a, p. 19). Populus sp. was among the overstory trees providing over 50 percent shade (Ferguson 2016a, p. 19). The site was located near an old mining road and hiking trail; the area did not burn in the 2005 Florida Fire (Ferguson 2016a, p. 19). The second group was on multiple aspects, from 1 to 9.5 m (3.3 to 31.2 ft) from an intermittent streambed in soil pockets and cracks of bare rock (Ferguson 2016a, p. 19). Populus sp., Salix sp. and Robinia sp. were among the overstory associates which provided over 50 percent shade (Ferguson 2016a, p. 19). Within close proximity of an old mining road and mining sites and hiking trail, the area did not burn in 2005 Florida Fire (Ferguson 2016a, p. 19).

Madera Canyon population Graptopetalum bartramii plants in the Madera Canyon subpopulation were found in multiple groups. One group was on west-southwest facing slopes of 45 to 80 degrees (Ferguson 2016a, p. 18). Associates included Platanus sp., which provided over 80 percent shade to the site (Ferguson 2016a, p. 18). Nearby disturbance included a trail and piping. A second group was found on a southwest facing slope of 30 to 90 degrees (Ferguson 2016a, p. 18). Associates here were Quercus sp. and Juniperus sp. providing over 50 percent shade to the site. This group was also near a trail. The stream channel contained rock debris following the 2005 Florida Fire (Ferguson 2016a, p. 18). A third group was on a south-southeast facing slope of 90 degrees that was 50 percent shaded by Pinus sp., Quercus sp., and canyon slopes (Ferguson 2014, p. 17). This group was in soil pockets and bare rock at 0.5 to 4.5 m (1.6 to 14.8 ft) from a flowing stream (Ferguson 2014, p. 17). A fourth group was found 1 to 4 m (3.3 to 13.1 ft) from a flowing stream on a south facing slope of 30 to 80 degrees (Ferguson 2014, p. 18). These plants were over 50 percent shaded by Platanus sp. and Quercus sp. as well as canyon walls; plants were in soil pockets and in bare rock (Ferguson 2014, p. 18). Two additional groups were found in similar habitat.

Squaw Gulch population Graptopetalum bartramii plants in the Squaw Gulch subpopulation were on thin soil on northwest facing rocky slopes of a seasonal drainage near the historical Viceroy Mine. Overstory included J. deppeana, P. discolor, and Q. emoryi, and Q. arizonica.

Sycamore Canyon population Graptopetalum bartramii plants in the Sycamore Canyon subpopulation were found in two groups. The first was 1 to 7 m (3.3 to 23 ft) from a dry streambed on a northwest facing slope of 45 to 90 degrees (Ferguson 2014, p. 22). Plants were in loose soil in pockets and bare rock, shaded 50 to 80 percent by Fraxinus sp. and Quercus sp. (Ferguson 2014, p. 22). The second group contained 23 dead and no living plants in 2013, yet contained 53 individuals in 2015, presumably returning from a seedbank (Ferguson 2017b, p. 11). It was 1 to 4 m (3.3 to 13.1 ft) from a dry streambed on northeast facing slopes of 45 to 90 degrees (Ferguson 2014 p. 24). Plants were in soil pockets and on bare rock with 30 to 50 percent shade from Juniperus sp. and canyon walls (Ferguson 2014, p. 25).

Temporal Gulch population Graptopetalum bartramii plants in the Temporal Gulch subpopulation were found on shady north facing slopes in sandy soil on bedrock.

Graptopetalum bartramii plants in the Upper Jones subpopulation were found in dense shade on south and north facing slopes. Plants were found in silt on bedrock, 1 to 15 m (3.3 to 49.2 ft) from a creek with water flowing. Only a small portion of the canyon contained suitable habitat.

Walker Canyon population Graptopetalum bartramii plants in the Big Casa Blanca Canyon subpopulation were found primarily on east facing slopes of a side canyon growing in colluvium and in cracks in the rock.

Graptopetalum bartramii in the Walker Canyon Basin subpopulation were found above a tank on a southwest facing slope about 1 m from a drainage with Pinus and Quercus in the overstory.

WHETSTONE MOUNTAINS

Figure 5.2j. Whetstone Mountains known Graptopetalum bartramii populations.

There are two populations of G. bartramii in the Whetstone Mountains, both on National Forest Service lands: Death Trap Canyon and French Joe Canyon (Figure 5.2j). In total the Whetstone Mountains are known to support 222 adult and 3 seedlings of G. bartramii plants (Table 5.11).

Population Subpopulations Year Adults Seedlings Dead Groups Notes Death Trap Death Trap Springs 2013 35 0 5 1 Canyon (HDMS-EO-40) 2015 135 0 3 2 French Joe French Joe Canyon 2016 87 3 7 2 Canyon (HDMS-EO-55) Table 5.11 History of Whetstone Mountains Graptopetalum bartramii population, subpopulation, and group occurrences.

Death Trap Canyon population Graptopetalum bartramii plants in the Death Trap Canyon subpopulation were found 1 to 5 m (3.3 to 16.4 ft) from a flowing stream on a northeast facing slope of 30 to 80 percent (Ferguson 2014, p. 30). The plants were in soil pockets and bare rock near a spring in Quercus woodland, with overstory including Fraxinus velutina (velvet ash) and J. deppeana. Migrant trails and trash were seen in the area (Ferguson 2014, p. 30).

French Joe Canyon population Graptopetalum bartramii plants in the French Joe Canyon subpopulation were found in two locations, both in soil and silt on bedrock, the first at 10 m (32.8 ft) from a creek with a small amount of water, the second from 2 to 10 m (6.6 to 32.8 ft) from this creek. Overstory associates included Q. emoryi, J. deppeana, and P. discolor; researchers also noted the invasive nonnative grass E. lehmanniana nearby (Heritage Database Management System September 2017).

Mexico There are three mountain ranges with G. bartramii in Mexico. We have no historical or current population size estimates or habitat details on these three sites. We assume there is only one population per sky island mountain range. We further assume that these populations are extant, but with few individuals, since we do not have data that indicate otherwise.

SIERRA LAS AVIPAS This population is the Sierra Las Avipas in Nogales County, Sonora. This site was first discovered in 2002 and has not been revisited.

SIERRA LA ESCUADRA This population is near Colonia Pacheco in the Municipio Nuevo Casas Grandes, Chihuahua in the Sierra la Escuadra. The site was first reported in 1948 and has not been revisted.

SIERRA LA ESTANCIA This population is in Cuarenta Casas, northwest of Las Varas, Municipio Madera, Chihuahua in the Sierra la Estancia. It was first reported in 1980 and has not been revisited.

5.3. Current Population Resiliency

Methodology

In Chapter 3 we discussed the needs of a resilient population and identified and described the population and habitat factors needed for a resilient G. bartramii population (see section 3.3.1 Population Resiliency). In Chapter 4 we described the stressors affecting populations of G. bartramii. In this section, we describe our methodology for assessing the resiliency of each population. We first define our understanding of what the various population resiliency levels (i.e. High, Moderate, Low, and Extirpated; Tables 5.12) are for this species. To describe the population resiliency levels, we used the six population and habitat factors identified above in section 3.3.1 Population Resiliency, as they are the primary factors influencing G. bartramii: number of subpopulations, abundance, recruitment, riparian elements, precipitation, and shade.

For each of the six population and habitat factors described in Chapter 3 we developed condition categories (High, Moderate, Low, and Extirpated) to assess the condition of each factor for each population (Table 5.13) in order to determine the overall population resiliency. Some factors rely on qualitative metrics while with others, where more data is available, we were able to develop quantitative metrics. We assigned a numerical value to the condition categories, High=3, Moderate=2, Low=1, and Extirpated =0, so we could calculate an overall score.

For some populations the specific information was not available; however, using our best professional judgement we made assumptions to complete our analysis based on what we do know about this species and similar species, habitat conditions, and the data reported. Information regarding the habitat condition in roughly 23 percent of the populations is unavailable. Systematic, regular surveys have not been conducted throughout the full range of this species; however, many surveys were conducted between 2014 and 2017, which covered most sky island mountain ranges. Survey information within and among populations varies in timing, data collected, and surveyor. To assess abundance for each population we used the most recent survey data available.

High (Good) Moderate Low Extirpated A population with high A population with A population with low A population with resilience is where moderate resilience is resilience is where no resiliency is one abundance is high, the where abundance is abundance is low, the that might be number of moderate, the number of number of subpopulations extirpated subpopulations is high subpopulations is moderate is limited to one and completely, either and spatially dispersed and spatial distribution is spatial distribution is physically or with multiple groupings; limited with few limited; seed production functionally seed production is high, groupings; seed production is low, mortality exceeds because so few recruitment is such that is moderate, recruitment recruitment such that the individuals are the population remains and mortality are equal population is declining; present that stable or increases; and such that the population ability to withstand reproduction is able to withstand does not grow; ability to stochastic events or unlikely (e.g. little stochastic events or withstand stochastic events recover from stochastic to no cross recover to current or or recover from stochastic events is unlikely due to pollination, better condition from events is limited due to low abundance and flowering, seed stochastic events from low abundance and recruitment and limited production, or seedbank; with abundant recruitment and reduced seedbank; with limited genetic exchange). suitable habitat. seedbank; with some suitable habitat. suitable habitat. Table 5.12. Population resiliency category definition for Graptopetalum bartramii.

To assess recruitment we used the number of seedlings and dead plants reported in the most recent survey data that reported abundance. In addition, we used the most recent survey data that reported abundance for the number of subpopulations. There were three populations that did not have recruitment data and one population that did not have subpopulation data. In these instances, we used abundance as an indicator of recruitment and number of subpopulations. We made the assumption that if abundance was low, then recruitment and the number of subpopulations was low.

To assess riparian elements, we considered the distance from water and/or riparian vegetation as an indicator of water nearby. If we did not know the distance from water we assumed it was greater than 20–30 m (55–98 ft; i.e. Low). If we knew there was water nearby but did not know how close we assumed the water was of moderate distance away (between 10–20 m [33–66 ft]). If there were multiple distances to water provided, we took the average, as the data indicate that some plants are close and some are farther away, and an average score represents this condition.

To assess shade, we assumed Low condition if no data were available. If the amount of shade was not quantified, but we knew that some shade occurred, we assumed the shade was moderate. If there was a range of values provided for the amount of shade at the site overall, we used the lowest amount of shade reported, with the understanding that this would be the worst-case scenario for amount of shade being provided to G. bartramii at the site. We recognize that using the lowest range or the average of the factors could underestimate or overestimate the various factors which could then over or underestimate the condition of populations. However, this is the best available data.

Population Factors Habitat Factors Condition Winter Subpopulatio Categories Abundance Recruitment Riparian Elements (Oct-Mar) Shade ns Precipitation Water is within 10 m Number of Populations More than 12 inches of Three or more from individuals or Overstory cover adults in each contain more winter rain on average subpopulations riparian vegetation of Juniperus, High (3) population is seedlings (<1.5 during the past 5 years of plants / present indicating Quercus, Pinus or > 300 cm [0.6 in]) than as recorded at the population. subsurface water other is > 80%. individuals. dying individuals. nearest weather station. nearby.

Populations Water at or near the Between 6.1 and 12 Number of Overstory cover Two contain an equal surface (riparian inches of winter rain on adults in each of Juniperus, subpopulations number of vegetation present average during the past Moderate (2) population is Quercus, Pinus or of plants / seedlings (<1.5 indicating subsurface 5 years as recorded at 150 to 300 other is between population. cm [0.6 in]) to water) is within 10–20 the nearest weather individuals. 50 and 80%. dying individuals. m from individuals. station. Water at or near the Number of Populations 6 or fewer inches of Overstory cover One surface (riparian adults in each contain fewer winter rain on average of Juniperus, subpopulation vegetation present Low (1) population is seedlings (<1.5 during the past 5 years Quercus, Pinus or of plants / indicating subsurface < 150 cm [0.6 in]) than as recorded at the other is between population. water) is within 20–30 individuals. dying individuals. nearest weather station. 20 and 50%. m from individuals. Streambed near plants No Population is is dry and invaded by individuals made up primarily non-riparian plant No are found of dead and dying Overstory cover species indicating shift Ø subpopulations during individuals that has been of vegetation . surveys in do not produce removed. community and appropriate seed or no complete loss of microhabitat. individuals found. suitable habitat. Table 5.13. Condition categories for population factors and habitat factors used to create population resiliency.

We averaged all the condition category scores for each population to determine the overall resiliency score. To provide context for this score we established an overall resiliency scale from 0 to 3 to communicate our understanding of the overall condition of each population (Table 5.14). To determine the overall resiliency scale we first determined the highest score attainable (3) and the lowest score attainable (0). Within this range, we established four overall resiliency levels based on the number of population and habitat factors in the condition categories as shown in Table 5.14. Appendix 2 provides the ranking of each population and habitat factor for current condition.

Overall Resiliency Extirpated 0–0.48 Low .5–1.49 Moderate 1.5–2.49 High 2.5–3 Table 5.14. Overall resiliency scale with scoring.

Table 5.15 below indicates the current condition of each population for each sky island mountain range. We consider the Stronghold Canyon West population of the Dragoon Mountains to be the only population currently in High condition due to the abundance of individuals and groups of plants, good recruitment, and moderate to high habitat condition factors. The Squaw Gulch population in the Santa Rita Mountains has the lowest condition. Figure 5.3 depicts the species’ current condition across the United States portion of the range. One population (3 percent) is in High condition, 22 populations (67 percent) are in Moderate condition, 8 populations (24 percent) are in Low condition, and 2 populations (6 percent) have been extirpated.

Mar) Mar) Current Sky Island Subpopulation - Condition Shade Winter Oct Riparian Riparian Elements ( Abundance Precipitation Recruitment Subpopulation

Baboquivari Brown Canyon 1 1 3 2 2 2 Moderate Thomas Canyon 1 1 1 2 2 2 Low Chiricahua Echo Canyon 3 2 1 2 1 2 Moderate Indian Creek 0 0 0 0 0 0 Extirpated Dragoon Carlink Canyon 0 0 0 0 0 0 Extirpated Jordan Canyon 1 3 3 3 2 2 Moderate Sheephead 1 1 3 1 2 2 Moderate Slavin Gulch 1 1 1 2 2 2 Moderate Stronghold Canyon Moderate 2 2 3 2 2 2 East Stronghold Canyon High 3 3 3 2 2 2 West Empire Empire Mountains 0 0 0 0 0 0 Extirpated Mule Juniper Flat 1 3 1 2 1 1 Moderate Pajarito-Atascosa Alamo Canyon 1 1 3 3 1 2 Moderate Holden Canyon 1 1 3 1 1 1 Moderate Sycamore Canyon 3 2 3 3 1 2 Moderate Warsaw Canyon 1 1 3 3 1 1 Moderate Patagonia Alum Gulch 2 1 1 3 1 2 Moderate Chimenea-Madrona Rincon 1 1 3 1 1 2 Moderate Canyon Happy Valley North 0 0 0 0 0 0 Extirpated Happy Valley South 1 1 3 3 1 2 Moderate Santa Rita Adobe Canyon 1 1 3 1 2 2 Moderate Gardner Canyon 3 1 1 3 2 2 Moderate Josephine Canyon 2 1 3 3 2 2 Moderate Madera Canyon 1 1 1 3 2 2 Moderate Squaw Gulch 1 1 1 1 2 1 Low Sycamore Canyon 1 1 1 3 2 1 Moderate Temporal Gulch 2 1 3 2 2 3 Moderate Walker Canyon 2 1 1 3 2 2 Moderate Whetstone Deathtrap Canyon 1 1 1 3 1 2 Low French Joe Canyon 1 1 1 3 1 1 Low Sierra Las Avispas,Sonora Sierra Las Avispas ? ? ? ? ? ? Low Near Colonia ? ? ? Low Sierra La Escuadra, Chihuahua ? ? ? Pacheco Sierra La Estancia, Chihuahua Cuarenta Casas ? ? ? ? ? ? Low Table 5.15. Current resiliency of the 33 known Graptopetalum bartramii populations in the United States and Mexico; conditions within the three populations in northern Mexico are unknown.

Figure 5.3. Current condition of Graptopetalum bartramii populations in the United States. Current condition of three G. bartramii populations in Mexico are unknown, but presumed to be in Low condition.

5.4 Current Species Representation

No genetic studies have been conducted within or between the 33 populations of G. bartramii in southern Arizona and Mexico. However, we consider G. bartramii to have representation in the form of potential genetic exchange, at least between populations containing many subpopulations or many plants per subpopulation. For example, at the center of its range within the Dragoon Mountains, there are high numbers of individual G. bartramii per population and many nearby populations, and populations typically contain several subpopulations of plants allowing for genetic exchange through pollinators moving within different populations or subpopulations or through seeds moving through waterways. Another example is the Santa Rita and Pajarito- Atascosa Mountains, in which there are moderately high numbers of individuals per population and numerous populations per mountain range, also increasing potential for genetic exchange. In addition, because the plant occurs on multiple substrate types and at a range of elevations (1,067 to 2,042 m [3,500 to 6,700 ft]), there is likely some local adaptation and genetic differentiation among populations. This range in elevation provides a variety of climatic conditions for the species to inhabit. And lastly, in at least three locations (Flux Canyon, Sycamore Canyon, and Gardner Canyon populations), G. bartramii have been reported over many decades, indicating that these populations may have the genetic and environmental diversity to adapt to changing conditions.

Mountain ranges that have only one or two populations and/or only have one subpopulation per population, or low numbers of individuals per population with several miles between sky islands, such as the Baboquivari, Chiricahua, Whetstone, and Patagonia Mountains, may not be as genetically diverse because pollination or transport of seeds between populations may be very limited. Another example is the Mule Mountains population, which contains a large numbers of individuals, but for which there is only one subpopulation, and for which the nearest neighbor population is approximately 38 km (23.6 mi) away, similarly limiting gene flow through pollinators or seeds.

5.5 Current Species Redundancy

The G. bartramii populations in the United States and Mexico are naturally fragmented between mountain ranges. Currently, there are 29 extant G. bartramii populations spread across 12 different mountain ranges in southern Arizona and northern Mexico. Although this may imply redundancy across its range, note that 22 of the 29 extant populations contain fewer than 150 total individual plants. Further, 13 of the 29 populations have 50 individuals or less, and 4 populations have been extirpated. The Baboquivari, Chiricahua, Whetstone, and Patagonia Mountains have only one or two populations and/or only have one subpopulation per population, and low numbers of individuals per population. Further, these sky island mountain ranges are several miles away from the other sky island mountain ranges, so natural gene exchange or re- establishment following disturbance is unlikely. In addition, the Mule Mountains contain large number of G. bartramii individuals, but there is only one population and it is approximately 38 km (23.6 mi) away from the nearest population, making natural re-establishment of populations unlikely. In addition, this population is known to be contracting in size due to drying of habitat (The Nature Conservancy 1987, p. 2).

CHAPTER 6 – VIABILITY

We have considered what G. bartramii needs for viability, as well as the current condition of those needed resources (Chapters 2 and 3), and we reviewed the stressors that are driving the historical, current, and future conditions of the species (Chapter 4). We now consider what the species’ future conditions are likely to be. We apply our future forecasts to the concepts of resiliency, redundancy, and representation to assess the future viability of G. bartramii.

6.1 Introduction

The historical range of G. bartramii is presumed to be similar to the current range of the species in southern Arizona and northern Sonora and Chihuahua, Mexico. We are aware of four populations that have become extirpated in the United States in recent years, and a fifth which has contracted in size. In three instances, extirpation was associated with the drying of habitat which rendered it no longer suitable for the species to persist; we do not know the cause of extirpation in the fourth instance. There are 4 previously recorded populations that have not been revisited to attain current status; it is unknown if these populations are still extant. However we assume they are extant, because the best available information indicates presence and persistence at sites that maintain habitat conditions.

Recently established permanent plots, in addition to some resurveys of populations between 2013 and 2016 throughout the range of G. bartramii, indicate rapid change within groups of plants, in terms of recruitment, growth, and loss (Ferguson 2016a, p. 28; Ferguson 2017b, p. 8). Within two plots in the Sycamore Canyon population and one plot of the Madera Canyon population (Santa Rita Mountain Range), Ferguson found population size varied from season-to- season and year-to-year of the study (Ferguson 2017b, p. 8). For example, over the entirety of his two-year study (2015-2017), 188, 77, and 201 individuals were noted within these three plots. However, the highest number of individuals at a plot at any one time was 131, 55, and 97, respectively. Of particular importance are seedlings which may not survive into larger size classes, as well as plants that seem to have disappeared, perhaps dislodging and drying up with no trace or perhaps having been removed illegally. Table 6.1 below shows changes in population numbers in three plots during 6 visits between the fall of 2014 and spring of 2017. From these three glimpses into G. bartramii population increase and decrease, there is an overall net increase in population size, however within single years mortality outweighs seedling production (also see Ferguson 2014, pp. 6-39).

Fall 2014 Spring 2015 Fall 2015 Spring 2016 Fall 2016 Spring 2017 Net Box Canyon +20 / -12 +35 / -1 +90 / -24 +7 / -58 +15 / -22 +3 / -14 +39 A Box Canyon +17 / -0 No record +34 / -6 +5 / -4 +3 / -5 +0 / -5 +39 B Madera +78 / -28 +14 / -9 +39 / -47 +5 / -25 +19 - 15 +0 / -7 +24 Canyon Table 6.1. Gains and losses in individual G. bartramii determined from six visits to three permanent plots located in the Santa Rita Mountain Range. Data from Ferguson 2017a, Table 4.

The G. bartramii populations in the United States and Mexico are naturally fragmented between mountain ranges. Each population faces a variety of stressors at varying levels of risk into the future from natural and anthropogenic stressors, including the following: • loss of water in nearby drainages from mining and drought; • erosion, sedimentation, and burial from mining, livestock, wildlife, recreation trails and roads, cross border violators, and post-wildfire runoff; • trampling from humans, wildlife, and livestock • high severity wildfires ignited from recreationist, cross border violators, and lightning; • loss of shade from mining, drought, insect predation, flooding, and wildfire; • higher frequencies of freezing and flooding events from current and future climate change; • loss of seedlings, adults, and reproduction from current and future drought; • predation of individuals and shade trees; and • illegal collection.

Climate change has already begun to affect the regions of Arizona and Mexico where G. bartramii occurs, resulting in higher air temperatures, increased evapotranspiration, and changing precipitation patterns, such that water levels range-wide have already reached historical lows (CLIMAS 2014, entire). These low water levels put the populations at elevated risk of habitat loss due to the reduction of humidity and shade. For the G. bartramii populations with low numbers of individuals or low habitat condition, a single stochastic event such as a wildfire or drought could eliminate an entire population. We are currently aware of three populations that have become extirpated in recent years and a fourth population that was extirpated four unknown reasons. These impacts are heightened at the species level because the isolation of the populations prohibits natural recolonization between sky island mountain ranges. Smaller populations are also more vulnerable to smaller stressors such as trampling, erosion, collection, and freezing. If populations lose resiliency, they are more vulnerable to extirpation, with resulting losses in representation and redundancy.

6.1.1 Scenarios Assessment

Because we have significant uncertainty regarding: (1) how much climate will change in the future, which in turn will have an effect on rainfall and severity of future periods of drought and flooding; (2) the number of wildfires and mines that will occur in the future; (3) whether nonnative plants will be aggressively removed or allowed to spread; (4) whether upland forests will be managed for wildfire prevention and more natural low severity fires, or if larger more severe fires will burn; and 5) whether cross border violators will continue to have a significant presence in these mountain ranges or be reduced by border control activities, we have forecast what G. bartramii may have in terms of resiliency, redundancy, and representation under four plausible future scenarios. These future scenarios forecast the viability of G. bartramii over the next 10 and 40 years. We chose 10 years to evaluate what is likely to occur in the near term, and because this is two generations of the plant; and 40 years because this is within the range of available hydrological and climate change model forecasts, and it represents eight generations of the plant.

• The following sky island mountain ranges are more heavily impacted by cross border violators than other sky island ranges containing G. bartramii: Baboquivari, Chiricahua, Mule, Pajarito-Atascosa, Santa Rita, Patagonia, and Whetstone Mountains. Consequently, we assume a higher risk of wildfire to populations in these sky island mountain ranges. • Nonnative grasses have been reported with G. bartramii in two instances, at French Joe Canyon and Juniper Flat populations. Consequently, we assume an increased likelihood of wildfire occurrence within G. bartramii populations in the Mule and Whetstone sky island mountain ranges. • The following are sky island mountain ranges containing current and anticipated increased future recreation: the Chiricahua, Dragoon, Pajarito-Atascosa, Patagonia, Santa Rita Mountains. Consequently, we assume a higher risk of wildfire and disturbance of plants in populations within these sky islands mountain ranges. • Because fires have occurred recently in each of the 10 sky island mountain ranges in the United States, we assume all of sky island mountain ranges have a chance of fire from lightning ignition. When nonnative, recreation, or cross border violator presence is also a risk in a particular range, wildfire ignition probability increases.

We developed 4 scenarios incorporating the stressors that are ongoing or will occur in the future to consider the range of possible future conditions. For each scenario we describe the level of impact from the identified stressors that would occur in each population. All of the scenarios involve some degree of uncertainty; however, they present a range of realistic and plausible future conditions. Table 6.2 below summarizes the four scenarios. All scenarios consider impacts from mining, wildfire, and climate. In addition, effects on individual plants from multiple stressors are assessed, including livestock, recreation, trampling, predation, and collection. Populations with less than 50 individuals, as determined by natural breaks in data, are assumed to be less resilient to these individual effects. The Continuation scenario evaluates the condition of G. bartramii if there is no increase in risks to the populations relative to what exists today, while the other scenarios evaluate the response of the species to changes in those risks. The Conservation scenario takes into account realistically possible additional protective measures which may or may not happen. The Moderate increase in negative effects scenario is an increase in the risks to populations with changes in climate as projected in a lower (4.5) emissions scenario along with increases in other stressors, as detailed below. The High increase in negative effects scenario is a further increase in risks to populations, with changes in climate projected at a higher (8.5) emissions scenario, and with additional increases in other stressors, as detailed below.

Mining Altered Fire Individual Risks Climate Climate Conservation Activity Regime Effects Risk Water Lightning Reduction Dislodging Livestock Conservation described extraction, Recreation in from flooding Recreation actions Excavation, Cross border available events, Trampling implemented Burial, violators water* and Seedling Predation Shade Nonnative / or shade desiccation, Collection reduction plants Flowering halt, Shade removal Scenario 1 Ongoing or Number of Available Number and Applied to No new individuals, Continuation planned wildfires water and severity of populations subpopulations or continuing mining annually drought flooding <50 populations found. into the activities as increases at continue at events individuals. No augmentation of future of 2012 the same the same continues at existing (~20). rate as the level as in the past 10 populations, no last 10 years. the past 10 years. seed preservation, years. Emissions nonnatives Emissions 4.5. controlled, and 4.5. forest thinned. Scenario 2 Number of Number of Available Flooding Applied to Sites revisited and Conservation mining wildfires water events do not populations additional plants activities does not remains increase. <50 are located, sites does not increase stable. Emissions individuals. are augmented, or increase from from current Emissions 4.5. new sites are current rate. 4.5. established, condition. nonnatives controlled, and forest thinned. Scenario 3 1–3 new Number of Available Increases in Applied to No new individuals, Moderate mining wildfires water is flash flooding populations subpopulations, or increase in activities increases in reduced per 4.5 <50 populations found, negative (above the uplands. per 4.5 emissions individuals. and no effects 2012 emissions scenario. augmentation of number) are scenario. existing implemented populations, and/or nonnatives existing controlled, and mines forest thinned. expand. Scenario 4 >3 new Number of Available Increases in Applied to No new individuals, Major mining wildfires water is flash flooding populations subpopulations or increase in activities are increases in reduced per 8.5 <50 populations found, negative implemented uplands. per 8.5 emissions individuals. and no effects and/or emissions scenario. augmentation of existing scenario. existing mines populations, expand. nonnatives controlled, and forest thinned. Table 6.2 Future Scenario Descriptions. * Available water includes precipitation, soil moisture, humidity, surface water, aquifer recharge, reduction in riparian vegetation, and increased number of days without water.

To assess the level of impact from a stressor to a population, we developed Impact Categories (Table 6.3). For mining activities we assessed current and projected mining activity from the USFS (2012, entire) and determined the number of activities planned for each sky island mountain range. The range of current and projected mining activities varied from 1 to 10 per sky island mountain range. Based on this variability, we identified the impacts from a single mining activity, which could be exploration or a new mine, on a single sky island mountain to be low. Mining activities of 2 to 3 per sky island mountain range would have moderate impact, and greater than 3 mining activities would have high impact. For each scenario we do not know the location, type, or size of the mining activity; therefore, we are assuming an equal probability of mining activity impacting a G. bartramii population.

For altered wildfire regime, we assessed the probability of ignition sources (lighting, cross border violator warming and cooking fires, and recreationist activities such as shooting ranges, cigarettes, and camping) or fuel load conditions (nonnative vegetation), for each sky island. Based on the level of activity, we determined the level of impacts to a population. We assumed lightning had an equal probability of occurring in any of the mountain ranges, so all populations scored at least Low. From Figure 4.5 and researcher notes from the field, the following sky island mountain ranges were presumed to have higher occurrence of cross border violators: Baboquivari, Mule, Pajarito-Atascosa, Santa Rita, Patagonia, and Whetstone. The following sky island mountain ranges were presumed to have higher recreational activity: Chiricahua, Dragoon, Pajarito-Atascosa, Patagonia, and Santa Rita (Ferguson, pers. comm. April 7, 2017; Buckley, pers. comm. April 8, 2017). The following sky island mountain ranges were shown to have the nonnative E. lehmanniana within G. bartramii populations: Mule and Whetstone. Wildfire probability was given a 0.1 or low impact in all 10 sky island mountain ranges due to the possibility of a lightning strike ignition. In sky island mountain ranges with lightning plus an addition ignition source, or if nonnative plants were present, we assigned a moderate impact. If lightning plus two or more ignition sources or nonnative plants were present, we assigned a high impact.

Nonnative plant presence was noted in the Juniper Flat and French Joe populations, and this increased the probability of fire, thus increasing the score by 0.1. Cross border violator presence in the Baboquivari, Mule, Pajarito-Atascosa, Santa Rita, Patagonia, and Whetstone Mountains are anticipated to be higher than the other four sky island ranges. With this presence comes an increased risk of wildfire starts; thus populations in these ranges received an increasing score of 0.1. Recreation is anticipated to be highest in the Chiricahua, Dragoon, Pajarito-Atascosa, Patagonia, Santa Rita Mountains ranges, and these were also given an increased score of 0.1 due to potential wildfire starts from unattended campfires, etc. These factors were all considered regarding wildfire in the current and future scenarios presented herein.

To assess impacts from climate change and drought (water availability), we assessed the projected future emissions scenarios and the associated projected changes in precipitation and extreme events (flooding). We assumed that a low level of impact to water availability would be the continuing condition as it is currently occurring. For a moderate level of impact, we assumed the 4.5 emissions scenario; and for a high level of impact, we assumed the 8.5 emissions scenario (see 6 Climate Change and Drought above). We then deducted points from the current condition

accordingly. We combined all forms of water loss (e.g. from drought, mining groundwater extraction, evapotranspiration, etc.) into a single category for evaluation.

Based on climate change data, it is likely that the severity of storm events will increase, resulting in more runoff, more severe flooding events, and more erosion and sedimentation affecting populations, especially following wildfire events in the uplands. G. bartramii need crevices in solid bedrock or in shallow soil pockets on rock ledges and cliffs in deep, narrow canyons above normal floodlines to avoid seeds and plants being washed away during flood events. An increase in the flood frequency or intensity could result in an increase in the number of plants dislodged. We estimated that plants within 2 m (6.6 ft) of a streambed are highly likely to be removed during flooding events and therefore considered populations with plants occurring within 2 m (6.6 ft) to be in Low condition because they are more susceptible to flood events. Plants located between 2 and 10 m (6.6 and 32.8 ft) of a streambed are moderately likely to be removed (Moderate), and those over 10 m (32.8 ft) are highly unlikely to be removed from flooding event (High). We do not know the floodline for each of the streams that contain the species, however, we assume that the above breakdown in distance from the stream represents flooding potential.

Precipitation determinations were made based on Western Region Climate Center Total precipitation for the months of October-March (deemed most important for G. bartramii germination and growth) averaged over the previous five year period. Areas with one inch or less of precipitation during these winter months ranked Low, between 6.1 and 12 inches ranked Moderate, and over 12 inches ranked High. To determine High, Moderate, or Low ranks, 333 points (weather station total precipitation by month over period of record) were examined within the range of G. bartramii; of these, 35 received over 12 inches of winter precipitation, 148 received between 6.1 and 12 inches, and the remaining 150 received less than 12 inch of winter precipitation.

Populations with fewer than 50 individuals are assumed to be less resilient to effects to individuals because most, if not all, individuals could be impacted to the point of population extirpation. For a population of 31–50 individuals we assumed low impact, for 21–30 individuals we assumed a moderate impact, and for populations of fewer than 20 individuals we assumed high impact (Table 6.3).

Climate Altered Climate Level of Mining (distance to Wildfire (availabl Individual Effects Score impacts Activity stream- Regime e water) flooding) Populations with an Low 1 Lightning 4.5 >10 m -.1 abundance of 31–50. Lightning plus Populations with an Moderate 2–3 rec or ill or 4.5 2–10 m -.2 abundance of 21–30. nonnatives Lightning plus Populations with an High >3 ill plus rec, or 8.5 <2 m -.3 abundance of <20. nonnatives Table 6.3. Scores given for level of impacts to populations from mining activity (including water withdrawal, erosion, sedimentation, burial, and loss of shade), altered wildfire regime (including loss of shade, erosion, burial, and direct mortality), climate induced water reduction (including drying of microhabitat, reduction in shade, and

flooding), and individual effects (stressors that have larger impacts on small populations such as trampling, erosion, predation, and collection).

The Overall Resiliency categories are the same as those used for Current Condition. As we did with the current condition assessment, we averaged all of the condition category scores for each population for each scenario at the various time steps to determine the overall resiliency score. To provide context for this score, we established an overall resiliency scale from 0 to 3 to communicate our understanding of the overall future condition of each population (Table 5.14).

We examined the resiliency, representation, and redundancy of G. bartramii under each of these plausible scenarios (Table ES–1). Resiliency of G. bartramii populations depends on future availability of appropriate riparian elements, precipitation, shade, substrate, and pollinators. We expect the 30 known populations of G. bartramii to experience changes to these aspects of their habitat in different ways under the different scenarios. We projected the expected future resiliency, representation, and redundancy of G. bartramii based on the events that would occur under each scenario (Table ES–2). As a general rule, when a population condition reaches the Low category, the population is at greater risk of being lost to a single event such as a fire, drought, predation, or illegal collection. Appendix 2 provides the ranking of each population and habitat factor.

6.2. Scenario 1 – Continuation

Under the Continuation scenario, those factors that are having an influence on populations of G. bartramii continue on the same trajectory they are on now. Low levels of climate change are already occurring at all locations, leading to a small water flow reduction, reduced humidity in the microhabitat and increased wildfire in uplands. In this scenario, the following risks remain probable: wildfire, cross border violator traffic, flooding, and mining that can impact individual populations. Collection, trampling, freezing, predation, and human impacts also continue at current levels.

6.2.1. Resiliency, Representation, and Redundancy

As identified above, we consider G. bartramii to have representation in the form of genetic diversity due to the species having historically occurred in 33 populations spread across 13 southern Arizona and northern Mexico mountain ranges. At the 10 year-time step, no populations would be in High condition, 4 populations (12 percent) would be in Moderate condition, 23 populations (70 percent) would be in Low condition and more susceptible to loss, and 6 populations (18 percent) would be extirpated (Table 6.4). Under this scenario, redundancy would decrease from the current condition, because 2 additional populations would become extirpated. Further, the loss of the 6 populations would reduce connectivity. Overall, under the Continuation scenario, 13 populations would retain current condition levels, and the remaining populations would have reduced condition relative to current conditions (Figure 6.1; Table 6.4). The current condition for many populations is just barely within the Moderate category cutoff of greater than 1.5, therefore continuing current threats reduced these levels enough to put them into the Low condition class.

Scenario 1 – Sky Island Population Continuation Baboquivari Brown Canyon Low Thomas Canyon Low Chiricahua Echo Canyon Low Indian Creek Extirpated Dragoon Carlink Canyon Extirpated Jordan Canyon Moderate Sheephead Low Slavin Gulch Low Stronghold Canyon East Moderate Stronghold Canyon West Moderate Empire Empire Mountains Extirpated Mule Juniper Flat Low Pajarito-Atascosa Alamo Canyon Low Holden Canyon Extirpated Sycamore Canyon Moderate Warsaw Canyon Low Patagonia Alum Canyon Low Rincon Chimenea-Madrona Canyon Low Happy Valley North Extirpated Happy Valley South Low Santa Rita Adobe Canyon Low Gardner Canyon Low Josephine Canyon Low Madera Canyon Low Squaw Gulch Extirpated Sycamore Canyon Low Temporal Gulch Low Walker Canyon Low Whetstone Deathtrap Canyon Low French Joe Canyon Low Sierra Las Avispas, Sonora Sierra Las Avispas Low Sierra La Escuadra, Chihuahua Near Colonia Pacheco Low Sierra La Estancia, Chihuahua Cuarenta Casas Low Table 6.4. Graptopetalum bartramii population conditions under the Continuation scenario.

Figure 6.1. Condition of Graptopetalum bartramii under the Continuation scenario.

6.3. Scenario 2 – Conservation

The Conservation scenario provides an idea of the best possible condition over the next 40 years. Under the Conservation scenario, those factors that are having a negative influence on populations of G. bartramii continue at current rates. Low levels of climate change and drought are already occurring at all locations, leading to a small water flow reduction, reduced humidity in the microhabitat, and increased wildfire in uplands. In this scenario, the following risks remain probable: wildfire, cross border violator traffic, flooding, and mining. Collection, trampling, freezing, predation, and human impacts also continue at current levels. However, conservation measures, including sites revisited and additional plants are located, sites are augmented, or new sites are established are implemented, seeds are preserved, nonnatives are controlled, and forests are thinned.

6.3.1. Resiliency, Representation, and Redundancy

Because current stressors (e.g. drought, wildfire, nonnatives, etc.) remain in place, conservation measures (e.g. population augmentation, wildfire prevention, and nonnative control) may, at best, render current conditions static for 21 populations, including the 4 currently extirpated populations remaining extirpated. Under this scenario, no populations would be in High condition, 11 populations in Moderate condition (33 percent), 18 in Low condition (55 percent), and the 4 populations (12 percent) currently extirpated remain extirpated (Table 6.5). Overall, 12 populations (36 percent) would decrease from their current condition due to continuing threats at current levels, and no populations would increase in condition class (Figure 6.2; Table 6.5).

Sky Island Population Scenario 2 – Conservation Baboquivari Brown Canyon Moderate Thomas Canyon Low Chiricahua Echo Canyon Moderate Indian Creek Extirpated Dragoon Carlink Canyon Extirpated Jordan Canyon Moderate Sheephead Low Slavin Gulch Low Stronghold Canyon East Moderate Stronghold Canyon West Moderate Empire Empire Mountains Extirpated Mule Juniper Flat Low Pajarito-Atascosa Alamo Canyon Low Holden Canyon Low Sycamore Canyon Moderate Warsaw Canyon Low Patagonia Alum Canyon Low Rincon Chimenea-Madrona Canyon Low Happy Valley North Extirpated Happy Valley South Moderate

Sky Island Population Scenario 2 – Conservation Santa Rita Adobe Canyon Low Gardner Canyon Moderate Josephine Canyon Moderate Madera Canyon Low Squaw Gulch Low Sycamore Canyon Low Temporal Gulch Moderate Walker Canyon Moderate Whetstone Deathtrap Canyon Low French Joe Canyon Low Sierra Las Avispas, Sonora Sierra Las Avispas Low Sierra La Escuadra, Chihuahua Near Colonia Pacheco Low Sierra La Estancia, Chihuahua Cuarenta Casas Low Table 6.5. Graptopetalum bartramii population conditions under the Conservation scenario.

Figure 6.2. Condition of Graptopetalum bartramii under the Conservation scenario.

6.4. Scenario 3 – Moderate increase in effects

Under the Moderate increase in effects scenario, water flow reduction due to drought and groundwater extraction continues to reduce the humid microhabitat for this species. Cross border violator traffic continues and risk of catastrophic wildfire is high due to dry conditions, invasion of nonnatives in the uplands, and increased risk of fire starts from illegal activity, recreation, and natural causes. Mining impacts individuals in the Patagonia and Santa Rita Mountains. Collection, trampling, freezing, predation, and human impacts also continue at current or increased levels.

6.4.1. Resiliency, Representation, and Redundancy

As identified above, we consider G. bartramii to have representation in the form of genetic diversity due to the species having historically occurred in 33 populations spread across 13 southern Arizona and northern Mexico mountain ranges. Under the Moderate increase in effects scenario, no populations would be in High condition, 4 populations (12 percent) would remain in Moderate condition, 16 populations (52 percent) would be in Low condition, and 13 populations (36 percent) would be extirpated, further reducing population redundancy and connectivity (Table 6.6). Under the Moderate increase in effects scenario, because of the intensity of stressors discussed above, 22 populations would be reduced from their current condition (Figure 6.3; Table 6.6). We further assumed that in the Moderate increase in effects scenario, one of these small populations becomes extirpated; and in the Major increase in effects scenario, all three of these small populations becomes extirpated.

Sky Island Population Scenario 3 – Moderate Increase Baboquivari Brown Canyon Low Thomas Canyon Low Chiricahua Echo Canyon Low Indian Creek Extirpated Dragoon Carlink Canyon Extirpated Jordan Canyon Moderate Sheephead Low Slavin Gulch Low Stronghold Canyon East Moderate Stronghold Canyon West Moderate Empire Empire Mountains Extirpated Mule Juniper Flat Low Pajarito-Atascosa Alamo Canyon Low Holden Canyon Extirpated Sycamore Canyon Moderate Warsaw Canyon Extirpated Patagonia Alum Canyon Extirpated Rincon Chimenea-Madrona Canyon Low Happy Valley North Extirpated Happy Valley South Low Santa Rita Adobe Canyon Low

Sky Island Population Scenario 3 – Moderate Increase Gardner Canyon Low Josephine Canyon Low Madera Canyon Extirpated Squaw Gulch Extirpated Sycamore Canyon Extirpated Temporal Gulch Low Walker Canyon Extirpated Whetstone Deathtrap Canyon Low French Joe Canyon Extirpated Sierra Las Avispas, Sonora Sierra Las Avispas Low Sierra La Escuadra, Near Colonia Pacheco Extirpated Chihuahua Sierra La Estancia, Cuarenta Casas Low Chihuahua Table 6.6. Graptopetalum bartramii population conditions under the Moderate increase in effects scenario.

100

Figure 6.3. Condition of Graptopetalum bartramii under the Moderate increase in effects scenario.

101

6.5 Scenario 4 – Major increase in effects

Under the Major increase in effects scenario, there is a large reduction in water flow due to drought and groundwater extraction, which continues to reduce the humid microhabitat required by this species. Cross border violator traffic continues, and risk of catastrophic wildfire is high due to dry conditions, invasion of nonnatives in the uplands, and increased risk of fire starts from illegal activity, recreation, and natural causes. Collection, trampling, freezing, predation, and human impacts also continue at current or increased levels (Appendix 2).

6.5.1. Resiliency, Representation, and Redundancy

As identified above, we consider G. bartramii to have representation in the form of genetic diversity due to the species having historically occurred in 33 populations spread across 13 southern Arizona and northern Mexico mountain ranges. Under the Major increase in effects scenario, of the 33 populations, none would be in High condition, 1 population (3 percent) would be in Moderate condition, 11 (33 percent) would be in Low condition and more susceptible to loss, and 21 (64 percent) would be extirpated, further reducing redundancy and connectivity (Table 6.7). Because we cannot accurately evaluate the three populations in Mexico due to a lack of information, we assumed all three are currently in Low condition with few individuals. Under the Major increase in effects scenario, because of even higher intensity stressors in this scenario, the conditions of 28 of the 33 populations would be reduced from current conditions (Figure 6.4; Table 6.7 and Table 6.8). We further assumed that in the Moderate increase in effects scenario, one of these small populations becomes extirpated; and in the Major increase in effects scenario, all three of these small populations becomes extirpated.

Sky Island Population Scenario 4 – Major increase Baboquivari Brown Canyon Low Thomas Canyon Extirpated Chiricahua Echo Canyon Low Indian Creek Extirpated Dragoon Carlink Canyon Extirpated Jordan Canyon Low Sheephead Low Slavin Gulch Extirpated Stronghold Canyon East Low Stronghold Canyon West Moderate Empire Empire Mountains Extirpated Mule Juniper Flat Extirpated Pajarito-Atascosa Alamo Canyon Extirpated Holden Canyon Extirpated Sycamore Canyon Low Warsaw Canyon Extirpated Patagonia Alum Canyon Extirpated Rincon Chimenea-Madrona Canyon Low Happy Valley North Extirpated Happy Valley South Low

102

Sky Island Population Scenario 4 – Major increase Santa Rita Adobe Canyon Extirpated Gardner Canyon Low Josephine Canyon Low Madera Canyon Extirpated Squaw Gulch Extirpated Sycamore Canyon Extirpated Temporal Gulch Extirpated Walker Canyon Extirpated Whetstone Deathtrap Canyon Low French Joe Canyon Extirpated Sierra Las Avispas,Sonora Sierra Las Avispas Extirpated Sierra La Escuadra, Near Colonia Pacheco Extirpated Chihuahua Sierra La Estancia, Cuarenta Casas Extirpated Chihuahua Table 6.7. Graptopetalum bartramii population conditions under the Major increase in effects scenario.

103

Figure 6.4. Condition of Graptopetalum bartramii under the Major increase in effects scenario.

104

6.6. Status Assessment Summary

We used the best available information to forecast the likely future condition of G. bartramii. Our goal was to describe the viability of the species in a manner that will address the needs of the species in terms of resiliency, representation, and redundancy. We considered the possible future condition of the species, considering a range of potential scenarios that include important influences on the status of the species. Our results describe a range of possible conditions in terms of how many and where G. bartramii populations are likely to persist into the future (Table 6.8).

Scenario 3 Scenario 4 Current Scenario 1 – Scenario 2 – Sky Island Population – Moderate – Major Condition Continuation Conservation Effects Effects Baboquivari Brown Cyn Moderate Low Moderate Low Low Thomas Cyn Low Low Low Low Extirpated Chiricahua Echo Cyn Moderate Low Moderate Low Low Indian Extirpated Extirpated Extirpated Extirpated Extirpated Creek Dragoon Carlink Cyn Extirpated Extirpated Extirpated Extirpated Extirpated Jordan Cyn Moderate Moderate Moderate Moderate Low Sheephead Moderate Low Low Low Low Slavin Moderate Low Low Low Extirpated Gulch Stronghold Moderate Moderate Moderate Moderate Low Cyn E. Stronghold High Moderate Moderate Moderate Moderate Cyn W. Empire Empire Mts Extirpated Extirpated Extirpated Extirpated Extirpated Mule Juniper Flat Moderate Low Low Low Extirpated Pajarito- Alamo Cyn Moderate Low Low Low Extirpated Atascosa Holden Cyn Moderate Extirpated Low Extirpated Extirpated Sycamore Moderate Moderate Moderate Moderate Low Cyn Warsaw Moderate Low Low Extirpated Extirpated Cyn Patagonia Alum Cyn Moderate Low Low Extirpated Extirpated Chimenea- Rincon Moderate Low Low Low Low Madrona Happy Extirpated Extirpated Extirpated Extirpated Extirpated Valley N. Happy Moderate Low Moderate Low Low Valley S. Santa Rita Adobe Cyn Moderate Low Low Low Extirpated Gardner Moderate Low Moderate Low Low Cyn Josephine Moderate Low Moderate Low Low Cyn

105

Scenario 3 Scenario 4 Current Scenario 1 – Scenario 2 – Sky Island Population – Moderate – Major Condition Continuation Conservation Effects Effects Madera Cyn Moderate Low Low Extirpated Extirpated Squaw Low Extirpated Low Extirpated Extirpated Gulch Sycamore Moderate Low Low Extirpated Extirpated Cyn Temporal Moderate Low Moderate Low Extirpated Gulch Walker Cyn Moderate Low Moderate Extirpated Extirpated Deathtrap Whetstone Low Low Low Low Low Cyn French Joe Low Low Low Extirpated Extirpated Cyn Sierra Las Sierra Las Avispas, Low Low Low Low Extirpated Avispas Sonora Sierra La Near Escuadra, Colonia Low Low Low Extirpated Extirpated Chihuahua Pacheco Sierra La Cuarenta Estancia, Low Low Low Low Extirpated Casas Chihuahua Table 6.8. Graptopetalum bartramii population conditions under each scenario.

We projected the likelihood of each scenario occurring at10 and 40-year time steps based on the conditions of each scenario as defined below (Table 6.9). Based on IPCC projected climate change impacts and recent known impacts from drought, wildfire, and increased flooding events, near term impacts, i.e., at the 10-year time step, are moderate to very likely. Within Scenario 3 we assume increased confidence in climate change impacts over time based on IPCC projections, in which they state high confidence in emissions scenario 4.5 being exceeded. Conversely, we have higher confidence in the Continuation near term, and less confidence that Continuation conditions would remain in effect over 40 years. For Scenario 4 we have less confidence that emissions scenario 8.5 will occur based on IPCC (2014, entire). However, the IPCC is confident that the emissions will fall within the 4.5 to 8.5 range. The Conservation scenario is only somewhat to moderately likely over time, because climate change impacts are projected to increase, and no conservation measures that may be implemented could address drying of habitat.

We examined the resiliency, representation, and redundancy of G. bartramii under each of the four plausible scenarios (Table 6.8). Resiliency of G. bartramii populations depends on the presence of riparian elements; availability of precipitation, especially in the winter; presence of shade to maintain microclimate, presence of appropriate substrate, and presence of pollinators (though the last two factors were not analyzed due to inadequate available information). We expect the 29 extant G. bartramii populations to experience changes to these aspects of their habitat in different ways under the different scenarios. We projected the expected future

106

resiliency, representation, and redundancy of G. bartramii based on the events that would occur under each scenario.

Scenario 3 – Likelihood of Current Scenario 1 – Scenario 2 – Moderate Scenario 4 – Major Scenario Condition Continuation Conservation increase in increase in effects Occurring at: effects 10 years n/a Highly likely Unlikely Somewhat likely Somewhat likely 40 years n/a Somewhat likely Somewhat likely Moderately likely Moderately likely Table 6.9. Highly likely = we are 90% sure that this scenario will occur; Moderately likely = we are 70–90 % sure that this scenario will occur; Somewhat likely = we are 50–70% sure that this scenario will occur; Unlikely= we are less than 50% sure that this scenario will occur.

Under scenario 1 – Continuation, we would expect the viability of G. bartramii to be characterized by a loss of resiliency, representation, and redundancy at the level that is currently occurring. At the 10-year time step, no populations would be in High condition, 4 populations (12 percent) would be in Moderate condition, 23 populations (70 percent) would be in Low condition and more susceptible to loss, and 6 populations (18 percent) would be extirpated. Within Scenario 1 we assume impacts from drought, climate change, and other stressors continue as in the near past. We think this is highly likely to occur within the 10-year time step with decreasing likelihood at future timesteps. This is because based on climate change projections, emissions will increase, resulting in increased impacts to the species.

Under scenario 2 – Conservation, we would expect the viability of G. bartramii to be characterized by higher levels of resiliency, representation, and redundancy than it exhibits under the current condition. However, because current stressors remain in place, conservation measures would result in populations similar to those under current conditions at best: static at the 40-year time step, with no populations in High condition, 11 populations in Moderate condition (33 percent), 18 in Low condition (55 percent), and the 4 populations (12 percent) currently extirpated remain extirpated. Within Scenario 2 we assume impacts from drought, climate change, and other stressors continue as in Scenario 1 but with conservation measures addressing stressors to the species. The conservation measures that could address stressors include nonnative control, forest thinning, and prevention of human caused wildfire. However, because climate change impacts are projected to increase and no conservation measures could address drying of habitat, we think this Scenario is somewhat likely in the long term.

Under scenario 3 –Moderate Effects, we would expect the viability of G. bartramii to be characterized by lower levels of resiliency, representation, and redundancy than it has in the Continuation scenario. At the 40-year time step, no populations would be in High condition, 4 populations (12 percent) would remain in Moderate condition, 16 populations (52 percent) would be in Low condition, and 13 populations (36 percent) would be extirpated. Within Scenario 3 we assume increased confidence in climate change impacts based on IPCC projections, in which they state high confidence in emissions scenario 4.5 being exceeded. Based on IPCC’s projections, we think that in the long term this is moderately likely.

107

1 population (3 percent) would be in Moderate condition, 11 (33 percent) would be in Low condition and more susceptible to loss, and 21 (64 percent) would be extirpated. For Scenario 4 the IPCC confidence is low that emissions scenario 8.5 will occur, therefore the likelihood of Scenario 4 occurring is moderately likely in the long term. However, the IPCC is confident that the emissions will fall within the 4.5 and 8.5 range.

108

Literature Cited

Acevedo-Rosas, R., V. Sosa, and F. Lorea. 2004. Phylogenetic relationships and morphological patterns in Graptopetalum (Crassulaceae). Brittonia, 56(2):185-194.

Aguilar, R., L. Ashworth, L. Galetto, and M. Aizen. 2006. Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta-analysis. Ecology Letters, (2006) 9: 968–980.

Anable, M. 1990. Alien plant invasion in relation to site characteristics and disturbance: Eragrostis lehmanniana on the Santa Rita Experimental Range, Arizona, 1937-1989. M.S. Thesis, University of Arizona. 64 pp.

Anable, M., M. McClaran, and G. Ruyle. 1992. Spread of introduced Lehmann lovegrass Eragrostis lehmanniana Nees. in southern Arizona, USA. Biological Conservation 61: 181-188.

Archer, S. and K. Predick. 2008. Climate change and ecosystems of the southwestern United States. Rangelands 30(3):23-28.

Arizona Department of Agriculture. Protected Native Plants By Categories. https://agriculture.az.gov/protected-native-plants-categories. Accessed April 21, 2016.

Austin, D. 2014. E-mail correspondence from Dan Austin, Arizona Sonora Desert Museum, to Julie Crawford, United States Fish and Wildlife Service. November 14, 2012.

Bahre, C. 1991. A legacy of change; Historic human impact on vegetation of the Arizona borderlands. The University of Arizona Press. 231 pp.

Bowers, J. 2005. Effects of drought on shrub survival and longevity in the northern Sonoran Desert. Journal of the Torrey Botanical Society 32(3):421-431.

Brooks, M. and D. Pyke. 2002. Invasive plants and fire in the deserts of North America. pp. 1- 14. In: K. Galley and T. Wilson (editors). Proceedings of the Invasive Species Workshop: the Role of Fire in the Control and Spread of Invasive Species. Fire Conference 2000: the First National Congress on Fire Ecology, Prevention, and Management. Miscellaneous Publication No. 11, Tall Timbers Research Station, Tallahassee, FL.

Brown, D. 1982. 123.3 Madrean Evergreen Woodland. Desert Plants. Pp. 59-65.

Brown, Patrick T. and Ken Caldeira. 2017. Greater future global warming inferred from the Earths recent energy budget. Nature: 552, December 7, 2017.

Buckley, S. 2013. E-mail correspondence from Steve Buckley, National Park Service, to George Ferguson, biologist. December 17, 2013.

109

Buckley, S. 2017. E-mail correspondence from Steve Buckley, National Park Service, to Julie Crawford, United States Fish and Wildlife Service. April 8, 2017.

Buckley, S. 2017. E-mail correspondence from Steve Buckley, National Park Service, to Julie Crawford, United States Fish and Wildlife Service. May 9, 2017.

Carroll, C., J.A. Vucetich, M.P. Nelson, D.J. Rohlf, and M.K. Phillips. 2010. Geography and recovery under the United States Endangered Species Act. Conservation Biology 24:395-403.

Center for Biological Diversity. 2010. Petition to list two Arizona plants from the Sky Islands as threatened or endangered under the Endangered Species Act. July 7, 2010.

Center for Desert Archaeology. 2005. Feasibility study for the Santa Cruz Valley Heritage Area; chapter 4I – mining booms 1680 to present. April 2005. http://www.archaeologysouthwest.org/pdf/scvnha/chapter04_i.pdf Accessed February 23, 2017

Chibuike, G. and S. Obiora. 2014. Heavy metal polluted soils: Effect on plants and bioremediation methods. Applied and Environmental Soil Science. 12 pp.

Chung, J., K. Park, C. Park, S. Lee, M. Chung, and M. Chung. 2009. Contrasting levels of genetic diversity between the historically rare orchid Cypripedium japonicum and the historically common orchid Cypripedium macranthos in South Korea. Botanical Journal of the Linnean Society 160:119-129.

CLIMAS 2014. Southwest Climate Outlook July 2014. http://www.climas.arizona.edu/blog/ southwest-climate-outlook-july-2014 Accessed July 31, 2014.

Cohan, D. 2011. E-mail correspondences from Dan Cohan, United States Fish and Wildlife Service to Julie Crawford, United States Fish and Wildlife Service. May 16, 2011.

Cohan, D. 2013. E-mail correspondences from Dan Cohan, United States Fish and Wildlife Service to Julie Crawford, United States Fish and Wildlife Service. February 22, 2013.

Crawford, J. C. Wahren, S. Kyle, and W. Moir. 2001. Responses of exotic plant species to fires in Pinus ponderosa forests in northern Arizona. Journal of Vegetation Science 12:261-268.

Crimmins, M. and A. Comrie. 2004. Interactions between antecedent climate and wildfire variability across south-eastern Arizona. International Journal of Wildland Fire 13:455-466.

Davis, M. and R. Shaw. 2001. Range shifts and adaptive responses to quaternary climate change. Science. 292:673-679.

Duncan, D.K., E Fernandez, and C. McCasland. 2010. The U.S.-Mexico border and endangered species. Pp. 123 – 130 In: Halverson, W. C. Schwalbe, and C. Van Riper III. 2010. Southwestern Desert Resources. University of Arizona Press, Tucson, Arizona.

110

Eddleman, K. 2012. Bioaccumulation of heavy metals from soils to plants in watersheds contaminated by acid mine drainage in SE Arizona. M.S. Thesis, Department of Chemistry and Biochemistry, University of Arizona. 197 pp.

Ellstrand, N. and D. Elam. 1993. Population genetic consequences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics. 24:217-242.

Felger, R. 2017. E-mail correspondences from Richard Felger, University of Arizona, to Julie Crawford, United States Fish and Wildlife Service. June 10, 2017.

Ferguson, G. 2017a. Post Brown Fire Re-survey of two Graptopetalum bartramii subpopulations in Baboquivari Mts., Pima Co., Arizona in 2016. March, 2017. 8pp.

Ferguson, G. 2017b. Multi-year study monitoring two Graptopetalum bartramii (Bartram Stonecrop, Crassulaceae) populations in Pima and Santa Cruz Counties, Arizona. Draft Final Report September, 2017. 24 pp. plus Tables.

Ferguson, G. 2017c. Graptopetalum bartramii (Bartram stonecrop) Draft Species Status Assessment - review comments, November 30, 2017.

Ferguson, G. 2017. E-mail correspondences from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. June 9, 2017.

Ferguson, G. 2017. E-mail correspondences from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. May 19, 2017.

Ferguson, G. 2017. E-mail correspondences from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. April 22, 2017.

Ferguson, G. 2017. E-mail correspondences from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. April 7, 2017.

Ferguson, G. 2016a. Graptopetalum bartramii re-survey and evaluation for study sites in Arizona, 2015-2016; Selected populations on Coronado National Forest and adjacent lands. A report submitted to the United States Forest Service. 37 pp.

Ferguson, G. 2016b. Post Mule Ridge Fire Re-survey of two Graptopetalum bartramii subpopulations in Santa Cruz Co., Arizona in 2016. 4pp. Report to USFWS, October 2016.

Ferguson, G. 2016. Email from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. November 18, 2016.

Ferguson, G. 2016. Email from George Ferguson, biologist, to Julie Crawford, United States Fish and Wildlife Service. July 15, 2016.

111

Ferguson, G. 2015. Section 6 Proposal submitted 5-8-2015. Multi-year study monitoring two Graptopetalum bartramii (Bartram’s Stonecrop, Crassulaceae) populations in Pima and Santa Cruz Counties, Arizona. 8 p.

Ferguson, G. 2014. Graptopetalum bartramii habitat assessment and census inventory in Arizona, 2013-2014; selected populations on Coronado National Forest and adjacent lands. A report submitted to the United States Forest Service. Contract AG-8197-P-13-0029. 76 pp.

Fernandez, R. and J. Reynolds 2000. Potential growth and drought tolerance of eight desert grasses: lack of a trade-off? Oecologia, 123, 90–98.

Fire Management Information System. 2016. Wildfire Report. Buenos Aires National Wildlife Refuge, Brown Fire, Number KA3V, 2016. 2pp.

FireScape. 2016. http://www.azfirescape.org/home. Accessed September 5, 2016.

Flora of North America. 2008. Volume 8. Magnoliophyta: Paeoniaceae to Ericaceae. Available at http://www.efloras.org/florataxon.aspx?flora_id=1&taxon_id=114025

Garfin, G., A. Jardine, R. Merideth, M. Black, and S. LeRoy, eds. 2013. Assessment of Climate Change in the Southwest United States: A Report Prepared for the National Climate Assessment. A report by the Southwest Climate Alliance. Washington, DC: Island Press. Gaylord, M., T. Kolb, W. Pockman, J. Plaut, E. Yepez, A. Macalady, R. Pangle,and N. McDowell. 2013. Drought predisposes piñon–juniper woodlands to insect attacks and mortality. New Phytologist 198: 567–578.

Geiger, E. and G. McPherson. 2005. Response of semi-desert grasslands invaded by non-native grasses to altered disturbance regimes. Journal of Biogeography 32:895-902.

Griffis, K., J. Crawford, M. Wagner, and W. Moir. 2000. Understory response to management treatments in northern Arizona ponderosa pine forests. Forest Ecology and Management 146:239-245.

Hart, S., T. DeLuca, G. Newman, M. MacKenzie, and S. Boyle. 2005. Post-fire vegetative dynamics as drivers of microbial community structure and function in forest soils. Forest Ecology and Management 220:166-184.

Humane Borders. 2017. http://www.humaneborders.info/app/map.asp. Accessed March 30, 2017.

InciWeb Incident Information System. 2016. Brown Fire; https://inciweb.nwcg.gov/incident/ 4793/. Accessed 3-7-2017

Intergovernmental Panel on Climate Change. 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A.

112

Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp. https://www.ipcc.ch/report/ar5/syr/ Accessed 2018.

Intergovernmental Panel on Climate Change. 2013. Summary for Policymakers. In T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.), Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. 29 pp.

Karl, T., J. Melillo, and T. Peterson. 2009. Global Climate Change Impacts in the United States. Cambridge University Press New York, NY. Available online from: http://www.globalchange.gov/publications/reports/scientific-assessments/us-impacts/download- the-report.

Kearney, R. and R. Peebles. 1951. Arizona Flora. University of California Press. 1,085. pp.

Kelly, J. and M.W. Olsen. 2011. Problems and pests of agave, aloe, cactus, and yucca. University of Arizona College of Agriculture and Life Sciences Publications – Extension – gardening and home horticulture. http://arizona.openrepository.com/arizona/bitstream/ 10150/144789/1/az1399-2011.pdf. Accessed February 2012.

Kodra, E., K. Steinhaeuser, and A. Ganguly. 2011. Persisting cold extremes under 21st‐century warming scenarios. Geophysical research letters 38:L08705 5 pp.

Korb, J., N. Johnson, and W. Covington. 2004. Slash pile burning effects on soil biotic and chemical properties and plant establishment: recommendations for amelioration. Restoration Ecology 12(1):52-62.

Malusa, J. 2001. Surveys of the sensitive plants of the Coronado National Forest, Arizona. A report to the National Forest Service under contract 43-8197-0-0083. 126 pp.

Mathias, A., G. Niu, and X. Zeng. 2013. Modeling the Climate and Hydrological Controls of the Expansion of an Invasive Grass Over Southern Arizona. Poster presented at the Tenth Annual Symposium of Research Insights in Semiarid Ecosystems. University of Arizona, Tucson, Marley Building, Rm 230, Saturday, 12 October, 2013.

Mazzoni, J., L. Green, A. Phillips, B. Phillips. 1980. Outline for field investigations Endangered plants project, Escheveria bartramii, October 11, 1980.

Moran, R. 1994. Crassulaceae stonecrop family. Journal of the Arizona-Nevada Academy of Science 27:190-194.

National Park Service (NPS). 2016. Graptopetalum surveys in southeastern Arizona 2015 – 2016. Interagency rare species research report – Draft version -. By Ashlee Simpson, Steve Buckley, Marcus Jernigan, and Tristan Jamieson. 12 pp.

113

National Park Service (NPS). 2015. Rare species memo, Graptopetalum bartramii at Chiricahua National Monument. 4 pp.

National Park Service (NPS). 2014. Resource brief, Graptopetalum bartramii at Chiricahua National Monument. 4 pp.

O’Dea, M. 2007. Influence of mycotrophy on native and introduced grass regeneration in a semiarid grassland following burning. Restoration Ecology 15(1):149-155.

Overpeck, J., G. Garfin, A. Jardine, D. Busch, D. Cayan, M. Dettinger, E. Fleishman, A. Gershunov, G. MacDonald, K. Redmond, W. Travis, and B. Udall. 2012. Chapter 1: summary for decision makers. In: Garfin, G., A. Jardine, R. Merideth, M. Black, and J. Overpeck, eds., Assessment of Climate Change in the Southwest United States: a Technical Report Prepared for the United States National Climate Assessment, A report by the Southwest Climate Alliance, Southwest Climate Summit Draft, Tucson, AZ.

Parmesan, C. and G. Yohe. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature. 421:37-42.

Peterson, T. 2003. Project climate change effects on Rocky Mountain and Great Plains birds; generalities of biodiversity consequences. Global Change Biology. 9:647-655.

Phillips, A., B. Phillips, N. Brian, J. Mazzoni, and L. Green. 1982. Status report on Graptopetalum bartramii Rose. Report by the Museum of Northern Arizona for the Office of Endangered Species, Fish and Wildlife Service. 15 pp.

Rathcke, B. and E. Jules. 1993. Habitat fragmentation and plant-pollinator interactions. Current Science 65(3):273-277.

Rawoot, D. 2017. Email from D. Rawoot, Land and Water Protection Manager, The Nature Conservancy, to Julie Crawford, United States Fish and Wildlife Service. September 18, 2017.

Robinett, D. 1992. Lehmann lovegrass and drought in southern Arizona. Rangelands 14(2): 100-103.

Rose, J. 1926. Graptopetalum bartramii (Bartram’s stonecrop) Native of Arizona. pp. 1 - 2 In: Addisonia, Colored illustrations and popular descriptions of plants. Volume II Number I. March 1926. Published by the New York Botanical Garden. 154 pp.

Schussman, H., E. Geiger, T. Mau-Crimmins, and J. Ward. 2006. Spread and current potential distribution of an alien grass, Eragrostis lehmanniana Nees, in the southwestern USA: comparing historical data and ecological niche models. Diversity and Distribution 12:582-592.

114

Seager, R., M. Ting, I. Held, Y. Kushnir, J. Lu, G. Vecchi, H.-P. Huang, N. Harnik, A. Leetmaa, N. C. Lau, C. Li, J. Velex, and N. Naik. 2007. Model projections of an imminent transition to a more arid climate in southwestern North America. Science 316:1181–1184.

Sebesta, D. 2016. Email from D. Sebesta, former Coronado National Forest Biologist, to Julie Crawford, United States Fish and Wildlife Service. June 20, 2016.

Shaffer. B. and L. Stein. 2000. In: Stein, B., L. Kutner, and J. Adams. Precious heritage: The status of biodiversity in the United States. The Nature Conservancy and Association for Biodiversity Information. Oxford University Press.

Shohet, C. 1999. Graptopetalum bartramii Rose demographics and reproductive biology, including pollination and flowering phenology (Crassulaceae). M.S. Thesis. Arizona State University. May 1999.

Simpson, A. 2017. Email from Ashlee Simpson, National Park Service, to Julie Crawford, United States Fish and Wildlife Service. October 4, 2017.

Smithwick, E., M. Turner, M. Mack, and F. Chapin III. 2005. Post-fire soil N cycling in northern conifer forests affected by severe, stand-replacing wildfires. Ecosystems 8:1630181.

Southwest Environmental Information Network (SEINet) - Arizona Chapter. 2017. http//:swbiodiversity.org/seinet/index.php. Accessed November 18, 2016 and January 27, 2017.

Stephens, S., J. Agee, P. Fule, M. North, W. Romme, T. Swetnam, and M. Turner. 2013. Managing forests and fire in changing climates. Science 342:41-42.

Stokes CA; MacDonald GE; Adams CR; Langeland KA; Miller DL, 2011. Seed biology and ecology of natalgrass (Melinis repens). Weed Science, 59(4):527-532.

Swetnam, T. C. Baisan, and J. Kaib. In press. Forest fire histories of La Frontera: fire-scar reconstructions of fire regimes in the United States / Mexico borderlands. In Press. In: G. L. Webster and C. J. Bahre eds., Vegetation and Flora of La Frontera: Historic vegetation Change Along the United States/Mexico Boundary. University of New Mexico Press.

Tersey, D. 2017. Email from D. Tersey, Bureau of Land Management, to Julie Crawford, United States Fish and Wildlife Service. March 30, 2017.

The Nature Conservancy. 1987. Site Stewardship Summary Escondido Falls Preserve. 17 p. report.

United States Fish and Wildlife Service (FWS). 2015. USFWS Species Status Assessment Framework: An integrated analytical framework for conservation. Version 3.3, dated August 2015.

115

United States Fish and Wildlife Service (FWS). 2012. Endangered and threatened wildlife and plants; 90-day finding on a petition to list Graptopetalum bartramii (Bartram stonecrop) and Pectis imberbis (beardless chinch weed) as endangered or threatened and designate critical habitat. 77 FR 47352. 5 pp.

United States Fish and Wildlife Service (FWS). 2011. Reinitiation of formal consultation on SBInet Ajo-1 Tower Project, Ajo Area of Responsibility, U.S. Border Patrol, Tucson Sector, Arizona : Proposed construction, operation, and maintenance of a forward operating base. Organ Pipe Cactus National Monument, Pima County, Arizona. September 16, 2011.

United States Fish and Wildlife Service (FWS). 1980. 45 FR 82480. Review of plant taxa for listing as endangered or threatened species. December 15, 1980.

United States Forest Services (USFS). 2013a. Final Environmental Impact Statement for the Rosemont Copper Project, a Proposed Mining Operation, Coronado National Forest Pima County, Arizona. Volumes 2 and 3.

United States Forest Service (USFS). 2013b. Draft Programmatic Environmental Impact Statement for Revision of the Coronado National Forest Land and Resource Management Plan Cochise, Graham, Pima, Pinal, and Santa Cruz Counties, Arizona; Hidalgo County, New Mexico. https://www.fs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb5440356.pdf. Accessed February 24, 2017.

United States Forest Service (USFS). 2012. Coronado National Forest Mining Activity (10 September 2012). 2 pp. + map.

United States Forest Service (USFS). 2008. Biological Evaluation. Authorization of grazing on the Barboot, Big Bend, Boss, Bruno, Hunt Canyon, Lower Rucker, Pedregosa, and Rak Grazing Allotments Douglas Ranger District , Coronado National Forest . May 12, 2008. 16 pp.

United States Forest Service (USFS). 2005. Forest Service sensitive species that are not listed or proposed under the ESA. 95 pp.

United States Geological Survey. 2018. Geosciences and Environmental Change Science Center. https://rmgsc.cr.usgs.gov/outgoing/GeoMAC/historic_fire_data/ Accessed July 2017 and January 2018.

United States Geological Survey. 2017a. The Water Cycle - USGS Water Science School - Evapotranspiration - The Water Cycle. https://water.usgs.gov/edu/watercycleevapotranspiration.html Accessed August 9, 2017.

United States Geological Survey. 2017b. Monthly Water Balance Model Futures Portal. https://www.usgs.gov/news/new-tool-shows-historic-and-simulated-future-water-conditions-us. Accessed August 9, 2017.

Van Devender. 1981. Report on field researches for Coronado National Forest. 28 pp.

116

Venable, D. and J. Brown. 1987. The selective interactions of dispersal, dormancy, and seed size as adaptations for reducing risk in variable environments. The American Naturalist 131 (3):360-384.

Waser, N. M. Price, G. Casco, M. Diaz, A. Morales, and J. Solverson. 2017. Effects of road dust on the pollination and reproduction of wildflowers. International Journal of Plant Science. 178(2):85-93.

Westerling, A., H. Hidalgo, D. Cayan, and T. Swetnam. 2006. Warming and earlier spring increase Wester United States forest wildfire activity. Science 313: 940-943.

Western Region Climate Center. 2017. Western U. S. Climate Summaries – NOAA coop stations. Arizona. https://wrcc.dri.edu/summary/Climsmaz.html. Accessed August 1, 2017.

Westland Resources Inc. 2013. Biological Evaluation - Proposed Arizona Trail Reroute, Northeastern Foothills of the Santa Rita Mountains, Pima County, Arizona. January 16, 2013. Project No. 1049.14. 89 pp.

Wolf, S., B. Hartl, C. Carroll, M.C. Neel, and D.N. Greenwald. 2015. Beyond PVA: Why recovery under the Endangered Species Act is more than population viability. Bioscience doi: 10.1093/biosci/biu218 1-8.

Zhang, Y., M. Hernandez, E. Anson, M. Nearing, H. Wei, J. Stone, and P. Heilman. 2012. Modeling climate change effects on runoff and soil erosion in southeastern Arizona rangelands and implications for mitigation with conservation practices. Journal of Soil and Water Conservation 67(5):390-405.

117

APPENDIX 1 Details on wildfires in the 10 US sky island mountain ranges known to support Graptopetalum bartramii. Fire year, hectares / acres burned, name of the fire, sky island mountain range where fire occurred, and approximate distance to the nearest known G. bartramii population is depicted for entire years of 2007 through 2017.

~ Distance (km) to Year Hectares Acres Fire Name Sky Island nearest known GRBA 2007 3,031 7,489 Alombre Baboquivari 10 2007 129 320 Fresno Patagonia 5 2007 346 855 Mansfield Santa Rita 0.6 2007 629 1,555 San Antonio Patagonia 16 Pajarito- 2008 2,055 5,079 Alamo 0.0001 Atascosa Pajarito- 2008 123 303 Beehive 7.6 Atascosa 2008 817 2,020 Buck Chiricahua 42 2008 19 48 Castro Rincon 1.7 Pajarito- 2008 103 255 Corral 1.8 Atascosa 2008 234 577 Cumaro Rincon 10.6 2008 3,060 7,562 Distillery Rincon 5.1 2008 128 317 High Lonesome Chiricahua 39 2008 3,483 8,606 Jackwood Chiricahua 30 2008 1,569 3,876 Meadow Patagonia 7.7 Pajarito- 2008 457 1,130 Nuevo 5 Atascosa 2008 1,042 2,575 Solano Baboquivari 8 Pajarito- 2008 4,067 10,050 White Tank 5 Atascosa Pajarito- 2009 42 105 Eagle Lake 2.7 Atascosa 2009 9,593 23,705 Elkhorn Baboquivari 0.0001 2009 65 161 Guidani Whetstone 2.6 2009 1,077 2,661 Lochiel Patagonia 14 2009 336 831 Mayday Chiricahua 36 2009 301 745 Rattlesnake Chiricahua 13.2 2009 55 135 Slavin Gulch Dragoon 2 2009 986 2,437 Three Peaks Baboquivari 6.1 2010 2,401 5,934 Brushy Chiricahua 23 Pajarito- 2010 398 984 Fraguita 12 Atascosa 2010 1,490 3,681 Horseshoe Chiricahua 21

118

~ Distance (km) to Year Hectares Acres Fire Name Sky Island nearest known GRBA Pajarito- 2011 3,926 9,702 Bull 7 Atascosa Pajarito- 2011 16 40 Divot 1.5 Atascosa 2011 3,119 7,708 Duke Patagonia 7.2 Pajarito- 2011 353 872 El Camino 10.7 Atascosa 2011 737 1,822 Greaterville Santa Rita 2.8 2011 91,038 224,961 Horseshoe 2 Chiricahua 0.0001 Pajarito- 2011 9 22 Jackie 1.3 Atascosa 2011 19 46 Milepost 6 Rincon 11 Pajarito- 2011 28,736 71,009 Murphy Complex 0.0001 Atascosa 2011 530 1,309 North Tank Patagonia 16.5 Pajarito- 2011 2,248 5554 Pena 0.1 Atascosa Pajarito- 2011 110 273 Verde 10.8 Atascosa 2011 161 398 Wildcat Patagonia 4 Pajarito- 2012 42 103 Ajax 16.5 Atascosa 2012 10,195 25,193 Apache Pass Chiricahua 21.7 2012 180 445 Babo Baboquivari 2.8 2012 65 160 Blind Chiricahua 35.8 2012 60 149 Bruno Chiricahua 35.4 2012 677 1,672 Cottonwood Whetstone 3.4 2012 60 149 Empire Empire 9 2012 3,108 7,680 Fox Rincon 2.1 2012 176 434 Hilton Empire 1.3 2012 2 6 Milepost 6 (2) Rincon 10.6 2012 679 1,678 Montezuma Baboquivari 1.3 2012 26 65 Rain Whetstone 9 2012 6 16 Reddington Rincon 10.5 2012 7 18 Sabe Patagonia 15.5 Pajarito- 2013 21 51 Chino 9.8 Atascosa 2013 4 11 Granite Peak Whetstone 3.3 2013 8 21 Hog Santa Rita 0.0001 Pajarito- 2013 2 6 Jalisco 8.9 Atascosa 2013 11 27 Kino Springs Patagonia 13.3

119

~ Distance (km) to Year Hectares Acres Fire Name Sky Island nearest known GRBA Pajarito- 2013 4 9 Lesna 16.4 Atascosa 2013 36 88 Milepost 8 Rincon 12.6 Pajarito- 2013 80 198 Peck 10 Atascosa Pajarito- 2013 17 43 Schumaker 1.6 Atascosa 2013 4,367 10,791 Soldier Basin Patagonia 0.0001 2013 325 804 Sycamore Baboquivari 5.2 2013 247 610 W2 Patagonia 16.1 2013 34 83 Washington Patagonia 15.8 2014 12 30 Bruno Chiricahua 36 2014 6 16 Bull Springs Santa Rita 3.3 2014 437 1,079 Deerhead Rincon 1.8 2014 14 34 Jackalope Rincon 8 Pajarito- 2014 17 43 Jalisco (2) 10.4 Atascosa Pajarito- 2014 1 2 Jarillas 14 Atascosa 2014 8 19 Middlemarch Dragoon 4.8 2014 12 29 Redington (2) Rincon 13.7 2014 15 38 Saulsbury Chiricahua 15 Pajarito- 2014 9 23 Tres Bellotas 9.1 Atascosa 2014 5 12 Turkey Chiricahua 15.8 2015 53 130 Aliso Rincon 6 Pajarito- 2015 6 15 Austerlitz 6 Atascosa 2015 10 24 Chivo Rincon 11.5 2015 45 110 Chumero Patagonia 5.1 Pajarito- 2015 15 37 Dos Amigos 0.88 Atascosa 2015 1 2 Dragoon Dragoon 2 2015 8 19 French Joe Whetstone 2.3 2015 92 227 Gunsite Pass Santa Rita 2.4 2015 825 2,039 Oak Tree Santa Rita 5.4 2015 793 1,960 Ojo Blanco Santa Rita 8.7 2015 780 1,928 Spring Baboquivari 0.0001 Pajarito- 2016 63 155 Bartlett 0.92 Atascosa Pajarito- 2016 1,795 4,435 Black Peak 3.5 Atascosa 2016 15,400 6,232 Brown Baboquivari 0.0001

120

~ Distance (km) to Year Hectares Acres Fire Name Sky Island nearest known GRBA 2016 366 905 Chovlic Baboquivari 9.1 2016 12 30 Cienega Creek Empire 13.7 Pajarito- 2016 120 296 Cruz 7.9 Atascosa Pajarito- 2016 2,249 5,557 Cumero 14.7 Atascosa Pajarito- 2016 284 701 Dicks Peak 9.2 Atascosa 2016 8 21 Emigrant Chiricahua 14.9 2016 90 223 Fresnal Baboquivari 14.1 2016 66 163 Hilton Empire 8.7 2016 12 30 Jhus Canyon Chiricahua 5.6 Pajarito- 2016 3,082 7,616 La Sierra 0.001 Atascosa Pajarito- 2016 3,389 8,375 Mule Ridge 0.0001 Atascosa Pajarito- 2016 30 75 Penasco 0.5 Atascosa 2016 145 358 Portal Peak Chiricahua 23.8 2016 4 9 Precopia Rincon 13.2 2016 366 904 Racetrack Rincon 12.9 2016 87 215 Round Corral Whetstone 8 Pajarito- 2016 32 79 Sonny 4.7 Atascosa Pajarito- 2016 8 21 Sycamore 0.4 Atascosa Pajarito- 2016 5 13 Triangle 7 Atascosa Pajarito- 2016 186 459 Triple 2 5.2 Atascosa Pajarito- 2016 28 70 Walker 4.8 Atascosa Pajarito- 2016 125 308 Warsaw 2.2 Atascosa 2016 38 95 Whitetail Canyon Chiricahua 4.5 2017 657 266 Elkhorn Baboquivari 4.5 2017 3,091 1,251 Bowie Chiricahua 13.5 2017 1,438 582 Fife Chiricahua 12 2017 1,130 457 Rucker Chiricahua 24 2017 85 35 Stanford Chiricahua 24.3 2017 30 12.2 Dragoons Dragoon 2.2 2017 14982 6063 Lizard Dragoon 0 2017 21 8.7 Noonan Dragoon 3.8 2017 222 90 Oak Empire 5.6

121

~ Distance (km) to Year Hectares Acres Fire Name Sky Island nearest known GRBA 2017 1,748 707 Mulberry Empire 0.28 2017 23 9 Alamo Atascosa 9.8 2017 279 113 Arrieta Atascosa 18.5 2017 607 246 Lesna Peak Atascosa 16.8 2017 94 38 Lobo Atascosa 11.5 2017 1,178 477 Mink Atascosa 15.2 2017 247 100 Pena Atascosa 3.8 2017 2,261 915 Flying R Patagonia 7.6 2017 1,057 428 Page Creek Rincon 4.5 2017 157 64 Fagan Santa Rita 8 2017 13 5 Gardner Santa Rita 1.5 2017 109 44 Santa Rita Santa Rita 6.4 2017 4,736 1,917 Sawmill Santa Rita 0.5 2017 94 38 Squaw Gulch Santa Rita 1.5 2017 105 42 Temporal Santa Rita 5.3 2017 38 16 Whetstones Whetstones 9.4

122

APPENDIX 2

See Future Scenarios Spreadsheet

123