In the face of a rapidly changing climate, spring-fed habitats are increasingly vulnerable to numerous threats. Botanical inventories provide valuable information to assess the conservation value of desert springs, and can serve as indicators to document changing conditions, including the proportion of native vs. nonnative taxa, diversity of life forms present that influence structure and function of ecosystems, species persistence and longevity, and the proportion of taxa that are rare and sensitive to land use change. Here we evaluate plant species composition and richness within and between springs, and evaluate botanical diversity with respect to physical parameters including hydrology and geography. We find that desert springs collectively support a large proportion of plant diversity, or nearly 22% of the total vascular plant diversity known within the California desert in only 0.000005% of the total land area. The springs we sampled are highly dissimilar in plant species composition, thus, restoration and management activities likely need to be highly individualized and site specific. Monitoring and inventory programs can increase opportunities for restoration and protection by providing information to assess warning signs of habitat degradation, such as changing species composition and local extirpation of wetland-dependent species.
INTRODUCTION
Desert springs have been identified as high-priority ecosystems and targets to meet conservation goals because of their potential to serve as microrefugia in a changing climate (McLaughlin et al. 2017; Cartwright et al. 2020; Parker et al. 2021). These small and isolated water features often provide perennial water sources in an otherwise arid landscape, supporting the persistence of species under changing conditions (Bogan et al. 2014). Despite their relatively small size (e.g., Jack Spring, San Bernardino County, California, is 0.755 ha), desert springs have wide-ranging ecosystem influence at a landscape scale (Patten et al. 2008; McLaughlin et al. 2017; Cartwright et al. 2020). At a broad scale, desert springs provide respite from harsh conditions for wide-ranging migratory animals (e.g., large mammals and birds), but at a local scale they support narrowly endemic species with limited habitat connectivity and dispersal capacity (Sada et al. 2005).
In the face of a rapidly changing climate, spring-fed habitats of the Mojave and Sonoran deserts, USA, and their associated biota face numerous threats including groundwater extraction, water diversion, introduction of nonnative plant species, impacts from feral animals (e.g., horses and donkeys), cattle grazing, and wide-scale land use change such as habitat alteration and conversion (Sada et al. 2005; Patten et al. 2008; Davis et al. 2017). Land management practices, allocation of resources, and land use changes can significantly alter fragile spring-fed habitats, and substantially impact spring-associated biodiversity and the conservation value of springs (Parker et al. 2021). Specifically, groundwater extraction has the potential to severely impact desert springs that rely on aquifers in regional settings when the extraction rate exceeds the recharge rate, resulting in aquifer drawdown (Zdon et al. 2018), or if pumping is proximal to a spring in a local groundwater system. Furthermore, recharge rates are expected to decline as climate conditions in the American Southwest are expected to become hotter and drier with longer and more frequent drought events (Meixner et al. 2016).
In the Mojave and Sonoran deserts in California (hereinafter referred to as the California desert), springs are largely understudied and biological information, especially with respect to plant diversity, is poorly documented (Zdon et al. 2018; Parker et al. 2021). Yet, plants are foundation species, and especially important in desert environments, because they ameliorate temperature, reduce direct evaporation, and provide food and nesting materials for wildlife (de Grenade 2013; Glenn et al. 2013). Documenting plant diversity is a key component to assessing habitat health and function of fragile wetland ecosystems and can provide essential information on baseline conditions (McLaughlin et al. 2017; Parker et al. 2021). Plant species can serve as bioindicators of spring condition because their distribution is highly influenced by a variety of environmental and anthropogenic factors such as substrate, light, temperature, water availability, pH, and level of disturbance. Furthermore, perennial plant species are relatively easy to survey and document, although field-based surveys can be time consuming and require specialized taxonomic expertise (Palacios Mejia et al. 2021). Therefore, increasing the knowledge of plant diversity, combined with an assessment of threats at desert springs, can provide a useful framework to inform restoration and resource management for these critical habitats.
As part of a growing, interdisciplinary body of work that characterizes the conservation values and needs of desert springs, we examined botanical diversity, assessed threats, and characterized hydrological parameters at 48 springs distributed across the California desert (Palacios Mejia et al. 2021; Parker et al. 2021). The goals of this study are to assess plant species composition and richness within and between springs, and to evaluate botanical diversity of wetland plants with respect to physical parameters including hydrology and geography. Specifically, we addressed the following questions: (1) What are the floristic patterns within and between desert springs? (2) Is plant species richness correlated with spring size? (3) How does human use and disturbance influence botanical diversity?
METHODS
Study Area
The California desert is located in the southeastern corner of the state and spans an area of 103,500 km2 (Omernick 1987). This study is focused on 48 springs that occupy an area of approximately 35 km2, or less than or 0.000005% of the total landmass of the California desert (Table 1; Figures 1–2). The springs chosen for this study are a subset of 341 springs surveyed by Andy Zdon and Associates (2016) that occur on land managed by the Bureau of Land Management (BLM); we selected these 48 springs because they span a broad geography and capture documented hydrological variation (Figure 1). The springs are well dispersed across the California desert with an average distance of 166 km between springs and only seven pairwise distances between springs that are below 0.5 km ( Supplemental Data Table 2 (naar-43-01-04_s02.xlsx)).
Field Surveys
Prior to surveys, we examined herbarium specimens, literature, and records from the California Department of Fish and Wildlife Natural Diversity Database (CNDDB 2018) to identify historical records and sensitive plant species previously documented across the survey area. We queried the Consortium of California Herbaria (CCH2 2019) and SEINet (2019) to produce a list of 523 unique historical herbarium specimens and establish baseline documentation of the local vegetation. We collected an additional 524 herbarium specimens in the field and submitted them in the herbarium at California Botanic Garden (formerly Rancho Santa Ana Botanic Garden [acronym RSA]), bringing the total number of herbarium specimens evaluated in this study to 1047 ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)).
We conducted surveys in the fall of 2018 and summer of 2019, when most wetland plant species are reproductive and in the best condition for identification. We identified plants to the minimum rank possible (species, subspecies, or variety), and excluded observations of plants if they could only be identified to genus or family. We collected herbarium specimens to aid in identification of species and provide a verifiable record. We did not collect specimens if plants were not in flower or fruit and could not be confidently identified. We included 222 observations of plant taxa that could be confidently identified in the field, but were not in a suitable condition to collect a voucher specimen ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). We examined all specimens cited in the study ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)) and identified them using taxonomic keys and descriptions from several references including Baldwin et al. (2012), Jepson Flora Project (2022), and Flora of North America (FNA 2021). We standardized names using the Jepson eFlora revision 9 (2022) and verified all specimen identifications through comparison with annotated specimens in the herbarium at RSA. Plant taxa were categorized by Wetland Indicator Status according to the National List of Plants that Occur in Wetlands developed by the U.S. Army Corps of Engineers (2022). Lifeforms were categorized using Calflora (2022).
Analysis
Plant species occurring at only one spring (singletons) were omitted for Jaccard's similarity index calculation because they contain no shared species and the resulting pairwise calculations would be zero. To analyze the percentage of wetland versus upland taxa at desert springs, we categorized wetland taxa as those occasionally found in wetlands (facultative upland) to taxa that always occur in wetlands (obligate) and all categories in between including facultative and facultative wetland. Upland taxa only consisted of obligate upland taxa (U.S. Army Corps of Engineers 2022). The areal extent of each spring was delineated using “heads-up” digitizing into a polygon feature class within Esri's ArcGIS Desktop 10.8 using the Albers equal-area conic projection. Each spring was located from field documented coordinates (Table 1) and the associated vegetation and footprint visible in NAIP 2018 1 m aerial imagery was traced onscreen. The footprint was determined by consensus of the authors. The CALCULATE AREA tool was used to calculate the hectares of each polygon. The GENERATE NEAR TABLE tool was used to calculate the distances between each of the spring polygons, in meters.
RESULTS
Physical Setting
Sampled springs spanned a wide geography across the California desert and ranged in elevation from 150 m to 1859 m and in size from 0.005 ha to 12.80 ha (Table 1, Figure 1). Underlying bedrock types are variable and include intrusive igneous rocks, extrusive igneous rocks (i.e., volcanic rocks), carbonate rocks (limestone and dolomite), and various metamorphic rocks (Andy Zdon and Associates 2016). The majority of the springs (81%) had water expressed at the surface during our surveys, while nine springs had no detectable water across multiple surveys where hydrological data was gathered (Table 1). The springs we sampled included local springs that are fed primarily by precipitation from their immediate watershed and influenced by local conditions and seasonal patterns, and regional springs that are fed by more extensive groundwater aquifers and are susceptible to impacts from regional groundwater pumping (Zdon and Love 2020).
Table 1.
Name, location, elevation, and size for 48 sampled springs. Taxa = minimum-rank taxa, *% = the percentage of nonnative taxa, WL% = the percentage of wetland plants present, and W = water present (1) or absent (0).
Floristic Summary
Plant species richness ranged from 3 to 171 taxa (mean 16.4 ± 28.7 SD) at each of the 48 springs sampled ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). We recorded a total of 479 minimum-rank plant taxa, across 78 plant families (Tables 2–3), with individual taxa present at 1–23 springs ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). Big Morongo Canyon was found to be the most species-rich desert spring site with 171 taxa documented. Six other springs have greater than 50 taxa documented (Table 1). Only 23 taxa (4.2%) were present at 10 springs or more, and 18 of these were native ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). A total of 185 taxa (38.5%) were individually present only at a single spring ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). A high proportion of the taxa recorded (87%) are native to the California desert (Table 2). The 10 most common plant taxa occurred at 13 sites or more and had a high proportion of wetland taxa (60%) and nonnative taxa (36%; Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)).
The five most common plant families across all springs were Asteraceae (99 taxa), Poaceae (41 taxa), Boraginaceae (34 taxa), Chenopodiaceae (26 taxa), and Polygonaceae (21 taxa; Tables 2–3). Springs were highly dissimilar based on Jaccard's similarity index with pairwise values ranging from 0 to 0.4 (mean 0.06 ± 0.05 SD). Our results indicated that spring size did not predict species richness when we evaluate all 48 springs together (R2 = 0.048; Table 1). However, if we omit two outlier springs (Saline Marsh Spring and Salt Spring), R2 increases to 0.611, indicating a positive relationship between spring size and species richness (Pearson's R = 0.781, P < 0.00001).
Figure 1.
Location of 48 springs in the Mojave and Sonoran deserts. All springs are on land managed by the Bureau of Land Management (Table 1).
Upland taxa are a significant component to the overall plant species composition at the desert springs sampled ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx), Figure 3). Upland taxa comprise 61.7% (295 taxa) of the total species richness while wetland taxa comprise 38.3% (183 taxa) of the total species richness (Figure 3). Of the 33 obligate wetland species documented, four were nonnative (12%) and the majority were perennial herbs (51%) and graminoids (36%). Cyperaceae was the most species-rich family among obligate wetland taxa (27%; Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). In contrast, of the 296 upland plant taxa we documented, only 6.7% were nonnative taxa, and their life forms primarily consisted of annual herbs (54%), shrubs (19.8%), and perennial herbs (15.8%) ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). Asteraceae was the most species-rich family among upland taxa (21.6%).
Baseline Botanical Documentation
There are relatively few herbarium records documenting floristic diversity at the 48 desert springs prior to this study. The majority of the historical records (71%) were collected by contemporary botanists or botanists who were active in the last 20 y (e.g., Duncan Bell, Naomi Fraga, Sarah De Groot, George Helmkamp, Andrew Sanders, and Justin Wood). Two prominent botanists made herbarium collections at three springs prior to 1970. Annie Alexander collected at Morongo Canyon Springs and Saline Marsh Spring in 1941 and 1955, respectively. Ernest Twisselman collected plants from Mesquite Spring in the El Paso Mountains in 1964. Plant taxa documented by Annie Alexander at Morongo Canyon Springs and Saline Marsh Spring were observed at both sites during our surveys and are considered extant at these sites. Four native perennial wetland taxa documented by Ernest Twisselman at Mesquite Spring were not relocated: Baccharis salicifolia subsp. salicifolia (facultative wetland shrub, Asteraceae), Pluchea sericea (facultative wetland shrub, Asteraceae), Eleocharis parishii (facultative wetland graminoid, Cyperaceae), and Phragmites australis (facultative wetland graminoid, Poaceae).
Taxa of Conservation Concern
We documented the occurrence of seven plant taxa of conservation concern; three of these (42.8%) were wetland taxa (Figure 4, Table 4). One species of conservation concern, Chloropyron tecopense (Orobanchaceae), was once found at Borehole Spring, Inyo County, California. This is a facultative wetland species that was last documented near Borehole Spring in 2008 (CNDDB 2018). Since 2015, multiple surveys have not relocated this taxon at this specific location, and it is presumed extirpated near Borehole Spring. All other rare plant occurrences that have been documented within the study area are currently presumed extant.
Table 2.
Floristic numerical summary. Taxa of conservation concern includes plants included in the California Native Plant Society's Inventory of Rare Plants (CNPS 2022). Calflora (2022) was used to categorize life forms.
Nonnative Taxa
We identified 63 nonnative taxa, which constitutes 13% of the total floristic diversity in the study (Table 2). The two most frequently encountered plant taxa across the study were nonnative grasses (Poaceae; Bromus rubens and Polypogon monspeliensis). These were the only two taxa in the study to occur at 20 springs or more ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). Cynodon dactylon (Poaceae), Schismus barbatus (Poaceae), and Erodium cicutarium (Geraniaceae) were relatively frequently encountered, occurring at 13, 13, and 12 sites, respectively.
Washingtonia filifera (Arecaceae) is native to the California desert, but is treated as a nonnative species in this study because it is known to be introduced at the specific sites sampled (Jepson eFlora 2022). Pulicaria paludosa (Asteraceae) is native to Portugal and Spain and was first recorded as a naturalized nonnative species in California in 1946 (FNA 2021). It primarily occurs in coastal California, but has also spread to the desert, primarily occurring in the low Sonoran Desert from Palm Springs to the Whipple Mountains (CCH2 2019). Pulicaria paludosa is common along roadways, streambeds, and seasonal wetland habitats (Jepson eFlora 2022). We documented it for the first time at Halloran Springs in San Bernardino County (Fraga 6440B, UCR), which represents a ∼190 km range extension to the north and it is the first record of the species in the central Mojave Desert region.
Nine of the documented nonnative plant taxa are persisting from cultivation. These include Ailanthus altissima (Simaroubaceae), Morus nigra (Moraceae), Nelumbo lutea (Nelumbonaceae), Nereum oleander (Apocynaceae), Parkinsonia aculeata (Fabaceae), Robinia pseudoacacia (Fabaceae), Tamarix aphylla (Tamaricaceae), Ulmus pumila (Ulmaceae), and Washingtonia filifera (Arecaceae) (Table 3). Taxa persisting from cultivation comprise 14% of all nonnative taxa documented in the study, 67% are categorized as wetland plant taxa, and 89% are trees or large shrubs (Table 3).
Figure 3.
Relative percentage of upland versus wetland plant taxa. UPL = Upland, FAC = Facultative, FACU = Facultative upland, FACW = Facultative wetland, OBL = Obligate (U.S. Army Corps of Engineers 2022).
DISCUSSION
Species Richness
Botanical diversity documented within the study area accounts for nearly 22% of the total vascular plant diversity known within the California desert, with just 48 springs containing 479 of the ∼2200 vascular plant taxa known to occur in the region (André 2014; Jepson eFlora 2021). This level of diversity is extraordinary given that springs sampled make up less than 1% (or 0.000005%) of the total land area for the California desert. One of the largest springs surveyed, Morongo Canyon Springs, was also one of the most species-rich; however, spring size did not predict species richness across all sites. This may be due to two outlier springs that occupy a large area (Saline Marsh Spring at 12.8 ha and Salt Spring 8.3 ha), but have relatively low species richness (22 and 15 plant taxa, respectively; Table 1). These two outlier springs have 1.7 taxa/ha and 1.8 taxa/ha, respectively, while the 48 springs together have a mean of 395 taxa/ha and mode of 444 taxa/ha (Table 1). When we remove Salt Marsh Spring and Salt Spring from our analysis a positive relationship between spring size and species richness is recovered, indicating that springs with larger footprints have the capacity to support increased plant diversity. A physical parameter like high pH or high salinity may be a limiting factor for species richness, as observed at Saline Marsh Spring and Salt Spring, which exhibit high pH (8 and 8.21, respectively) and high salinity (7096 ppm at Salt Spring).
Pattern of Dissimilarity between Springs
The results of the field surveys document a large number of unique plant taxa occurring within desert springs (i.e., singletons), with nearly 40% (185 of 479) of the documented plant taxa recorded at only one site, indicating high beta diversity, and little overlap in plant species composition between sites ( Supplemental Data Table 1 (naar-43-01-04_s01.xlsx)). This could be influenced by a number of factors including relatively low species richness at a large proportion of sites, with nearly 50% of the springs having fewer than 20 taxa present (Table 1). Dissimilarity between springs may also be influenced by the wide dispersion of sampling locations across a broad geographic region, and the relatively high proportion of upland taxa. Upland plant taxa were primarily composed of annual herbs, which are ephemeral by nature, and their occurrence may be more influenced by local conditions (e.g., precipitation and temperature) than wetland taxa.
Table 3.
A checklist of all vascular taxa documented across 48 desert springs. Nonnative species are denoted with an asterisk (*). Life form (Calflora 2021) and wetland status are provided. UPL = Upland, FAC = Facultative, FACU = Facultative upland, FACW = Facultative wetland, OBL = Obligate (U.S. Army Corps of Engineers 2022).
Continued.
Continued.
Continued.
Continued.
Continued.
Continued.
Continued.
The most frequently encountered taxa were wetland and nonnative taxa; these are groups that are known to have high dispersal capacity (Soomers et al. 2013; Schöpke et al. 2019). Wetland plant taxa are frequently dispersed by wind and water, and wind dispersal has been shown to be an effective means for relatively long-distance dispersal (Soomers et al. 2013). Wind-dispersed taxa (e.g., Populus fremontii, Salix sp., and Typha sp.) may be especially important contributors to community assembly following significant disturbance (e.g., fire, grazing, or other habitat modification) because these fragmented habitats are not hydrologically connected by surface water (Soomers et al. 2013). Nonnative taxa are frequently readily dispersed and easily established (McKinney and La Sorte 2007). The most common nonnative species we documented were annual grasses that are known for their wide dispersal capacity and potential to invade and dominate habitats (Curtis and Bradley 2015; Brooks et al. 2016).
Taxa of Conservation Concern
Eight taxa of conservation concern were documented within the study area; three of these are wetland plant taxa. There are at least 23 plant taxa of conservation concern known to occur in wetland habitats in the California desert (CNPS 2022). The majority of these were not documented as a part of this study because they may be associated with spring outflow habitats or shallow groundwater (e.g., alkali wetland, marshes, and meadows), and may not be located at spring sources themselves. Rare plant taxa in California also tend to have highly limited and specific distributions (Thorne et al. 2009). Systematic surveys are needed to determine the status of rare plant taxa associated with desert springs in California because rare plants and their habitat are highly impacted by the same suite of disturbances that affect desert springs including groundwater pumping, habitat conversion, invasive species, cattle grazing, and feral animals (Fraga et al. 2021; Parker et al. 2021).
Table 4.
List of plant taxa of conservation concern, their associated conservation ranks (CNPS 2022), and the number of springs in which they were documented.
Chloropyron tecopense (Tecopa bird's beak) is presumed extirpated at Borehole Spring. Since 2008, discharge at Borehole Spring appears to have slowly decreased, although recreational use of the spring and other human activities have increased substantially during that time frame resulting in channel modifications and vehicle trespass (Partner Engineering and Science, Inc. 2020). This demonstrates the vulnerability of rare plant taxa to local extinction when their habitats are subject to hydrological change and other disturbance.
Species Persisting from Cultivation
We documented nine species persisting from cultivation. These are primarily large shrubs or trees that are known to occur in wetland environments. These taxa can have a disproportionate influence on spring habitats (Sala et al. 1996; Fleishman et al. 2003; Neale et al. 2011). For instance, Tamarix can increase water loss from wetland environments due to high rates of evapotranspiration from their high leaf surface area (Sala et al. 1996). Similar mechanisms could be occurring in large trees and shrubs planted at sites that have large water demands. Further, large, woody, nonnative species have a greater community composition effect because they can displace native species and modify the structure and composition of the associated flora via competition for resources such as water, space, and light (Fleishman et al. 2003). However, wildlife (e.g., native birds and mammals) may also utilize these large trees and shrubs persisting from cultivation for shelter, shade, and nesting sites, complicating recommended management actions such as nonnative species removal.
Evidence of Floristic Change
Herbarium specimen records provide valuable baseline data that can be used to evaluate floristic change at sites, especially relating to changing hydrology and water availability at desert springs. We detected floristic change at Mesquite Spring, which currently occupies a relatively small footprint (0.142 ha) and no longer has surface water present (Table 1). Four perennial native wetland taxa no longer occur at the site and are presumed extirpated due to changes in hydrological conditions. Historical records indicate that water flowed at the site and that perhaps wild grapes (Vitis sp.) and roses (Rosa sp.) once grew there (Parker et al. 2021). Our study documented the loss of four wetland taxa. The floristic data provided here serve as another temporal point to evaluate change in plant species composition through time, which will become increasingly important as the climate is expected to become hotter and drier.
Floristic Diversity as a Metric for Conservation Value
A botanical assessment of the 48 springs in this study is a first step toward evaluating their relative importance for biodiversity conservation in a changing climate. Species richness is frequently used as a metric for assessing conservation value and priorities (Fleishman et al. 2006). However, beyond species richness, a more complete evaluation of the site-level floristic diversity can provide additional criteria to evaluate conservation value. This may include the proportion of native versus nonnative taxa, diversity of life forms that influence structure and function of ecosystems, species persistence and longevity, and the proportion of taxa that are rare and sensitive to land use change.
The results from this floristic inventory have several important implications for assessment of the conservation value of desert springs, and for land management and restoration activities. First, these ecosystems collectively support a large proportion of plant diversity in the California desert. They are a valuable resource for the conservation of landscape-scale plant diversity because they serve as reservoirs and refugia. Second, rare plant taxa are vulnerable to local extirpation, especially due to hydrological change. Third, the high beta diversity and dissimilarity between spring sites indicates that each spring represents a unique assemblage of plant species. Thus, restoration and management activities at California desert springs likely need to be highly individualized and site specific. Finally, to maximize the potential for desert springs to serve as refugia under changing climate conditions, inventory and monitoring are essential to recognize warning signs, including changing species composition and local extirpation of wetland-dependent species. Protection from non-climate threats such as water diversion and groundwater pumping, disturbance from feral animals, grazing, and habitat conversion are key to support long-term conservation of these life-sustaining ecosystems.
ACKNOWLEDGMENTS
We thank Duncan Bell, Erin Berkowitz, Wendy Boes, Sarah De Groot, Joy England, Carlos Garcia, Joselyn Gonzalez, LeRoy Gross, Nina House, Maria Jesus, Tiffany Larrabee, My-Lan Le, Anthony Perez, Mariana Rodriguez, Kim Schaefer, Kristy Snyder, Alejandra Soto, Byron Valdez, Rachel Wing, and Liz Womack for their assistance with field work. We thank Christina Lund and Chris Otahal at the Bureau of Land Management for assisting with permits. Patrick Donnelly provided comments on an early draft of this manuscript. This work was funded by The Nature Conservancy and California Botanic Garden.