R E S E A R C H A R T I C L E ABSTRACT: In support of natural resource agencies in Canada, the United States, and Mexico, we report on a series of component analyses and an updated Landscape Conservation Design for temperate • grassland conservation. We targeted 12 major grassland ecosystem types that occur across the Great Plains and Chihuahuan Desert regions. Component analyses included (1) documenting long-term trends in extent by grassland type, (2) identifying species of concern associated with the major grassland types, Continent- (3) documenting current protected areas including each grassland type, (4) assessing landscape intactness and connectivity among grassland areas, and (5) identifying Grassland Potential Conservation Areas Scale Landscape (GPCAs) to advance grassland conservation. Most severe declines in grassland extent have occurred in tallgrass prairie types, followed by mixed-grass, shortgrass, and semi-desert grasslands. Similar trends Conservation by type were documented for landscape intactness and connectivity. Some 174 species of vertebrates, invertebrates, and plants considered by NatureServe as critically imperiled, imperiled, or vulnerable are Design for strongly associated with these grassland types, and 103 are listed under protective legislation in one or more countries. Just 1.2% of historic extent for all types combined is currently found within designated Temperate protected areas. A total of 177 GPCAs were identified to represent grassland type diversity in areas least likely to conflict with other land uses. Within identified GPCAs, type-specific representation varied from Grasslands of the a low of just 1% of historic extent for Texas Blackland Tallgrass Prairie to a high of 27% for Western Great Plains and Great Plains Sand Prairie. Combined across all 12 grassland types, 15% of historic extent is represented. Chihuahuan Desert Index terms: Aichi Target 11, conservation land investments, Landscape Conservation Design, long-term trends in extent, temperate grassland types
Patrick J. Comer1,5 INTRODUCTION the Great Plains to Mexico’s Chihuahuan 1Ecology Department Desert grasslands have declined by almost NatureServe Due to land use intensification (Hoekstra 70% since 1970. Among the diverse her- 1680 38th Street Suite 120 et al. 2005) and projected impacts of cli- bivores of the grasslands, the black-tailed Boulder, CO 80301 mate change (Sala et al. 2000), temperate prairie dog (Cynomys ludovicianus Ord), grasslands are considered among the most often considered a “keystone” species in Jon C. Hak1 threatened biomes worldwide. North Amer- these grasslands, has experienced dramatic 2 Kelly Kindscher ican prairies and desert grasslands have population decline due to multiple factors Esteban Muldavin3 sustained extensive loss to land conversion including use of poison, habitat conversion, and degradation since the 1800s due to and sylvatic plague, and was recently con- Jason Singhurst4 agricultural conversion, overgrazing, and sidered for listing under the Endangered infrastructure development (Buffington and Species Act (Hoogland 2013). 2Kansas Biological Survey Herbel 1965). Those grasslands, extending University of Kansas from the Canadian prairie provinces of Regional patterns in climate (from cool and 2101 Constant Avenue Manitoba, Saskatchewan, and Alberta, moist to hot and dry) and soils explain much Lawrence, KS 66047 south across the US Great Plains and of the natural variation in major temperate through the Chihuahuan Desert of Northern grassland types. Tallgrass prairies occur 3Natural Heritage New Mexico Mexico, have experienced similar declines in a north–south belt from Manitoba to Biology Department to others found worldwide. Significant east Texas. Mixed-grass prairies dominate MSC03 2020 1 University of New portions of the vast trinational area (3M the Canadian prairies, extending south Mexico km2) are considered a “breadbasket of the throughout the central Great Plains to Albuquerque, NM 87131 world” with some of the most productive central Texas. Both shortgrass and sand and intensively cultivated croplands and prairies, occurring on loamy and sandy 4Texas Parks and Wildlife Department pasture lands (Gauthier et al. 2003). soils, respectively, predominate the west- 4200 Smith School Road ern and southern Great Plains. Finally, Austin, TX 78744 Wildlife has paid a steep price for this semi-desert grasslands occur throughout agricultural output. Grassland birds have the Chihuahuan Desert in southern Arizona, seen sharp declines with increasingly frag- New Mexico, and west Texas, and extend • mented and converted habitat (With et al. south and east to San Luis Potosí and 2008; Hill et al. 2014), as have many other Nuevo León, Mexico (Figure 1). 5 Corresponding author: grassland-dependent plants and animals. [email protected]; (703) 797- For example, the grassland avian commu- Patterns in habitat loss vary among grass- 4802 nity is among the highest conservation con- lands and are also related to climate and Natural Areas Journal 38:196–211 cern due to rapidly declining populations soils, which influence agricultural poten- (NABCI 2016). Species that migrate from tial. By combining agricultural potential
196 Natural Areas Journal Volume 38 (2), 2018 Figure 1. Approximate historical extent of 12 major temperate grassland types.
Volume 38 (2), 2018 Natural Areas Journal 197 with distribution of infrastructure and intact habitat, establishing redundancy Gap Analysis Program (https://gapanalysis. access to markets, one can identify major where needed, ensuring fine-scale and usgs.gov/gaplandcover/; see type descrip- trends of grassland conversion over the past broad-scale connectivity for species and tions at http://explorer.natureserve.org/). century. More recent trends in agricultural processes, providing refugia for species Recent investments in continental-scale intensification of the Chihuahuan Desert from rapid landscape change, and restoring mapping have identified 12 major grassland (Pool et al. 2014), biofuel production targeted habitats and species. The project types across this trinational area. (Wright and Wimberly 2013), and energy was designed to demonstrate the capacity development (Pruitt et al. 2009; Fargione of available data and analytical tools and Within the United States, LANDFIRE has et al. 2012) have introduced pockets of approaches to identify priority conservation produced map products for “biophysical rapid change across the region (Drummond areas to focus individual decisions and setting” of major upland ecological system et al. 2012). conservation investments by agencies, and type—effectively an estimate of potential/ potential coordinated efforts. Below we historical location and extent—using in- Prevailing patterns of land ownership shape report on a series of component analyses ductive modeling that utilizes field-based conservation priorities and actions in these leading to an updated Landscape Conser- observations of each type and correlates temperate grasslands. Privately held lands vation Design (LCD) identifying Grassland these with mapped information on cli- prevail throughout the Canadian and US Potential Conservation Areas (GPCAs) as mate, landform, and soil (Rollins 2009). prairies, while communal land ownership focal areas for conservation action. NatureServe then applied the same meth- is more frequently encountered in Mexico. ods with data extending across temperate Private ownership predominates where Component analyses included: Canada and south across Mexico (Comer agricultural productivity remains high, and (a) documenting long-term trends in extent et al. unpub. data) to provide a complete so historic triggers for concentrated public of 12 major temperate grassland types; distribution for each major grassland type ownership (e.g., tax-reversion of lands (b) identifying species of concern associ- at spatial resolutions of approximately 5-ha during the Great Depression) were limited ated with the major grassland types; minimum map unit. Using these estimates to shortgrass prairie and desert grasslands (c) documenting current protected areas of historical extent, the 12 most extensive (Gauthier et al. 2003; Samson et al. 2004). including each grassland type; grassland types were identified and served Given these ownership patterns, conserva- (d) assessing landscape intactness and as the focus for this analysis. Figure 1 tion largely takes place on lands controlled connectivity among grassland areas; and depicts predicted historical extent of each by private individuals, corporations, or (e) complete landscape conservation of the 12 major temperate grassland types. communal land stewards. Consequently, a design to identify GPCAs to advance These grassland types range in estimated mix of public land acquisition, government grassland resiliency. historical extent from a high of nearly subsidies for private conservation action, 630,000 km2 for Northwest Great Plains conservation easements, and multi-stake- Brief summaries of methods and results are Mixed-grass Prairie, occurring from Ne- holder partnerships (Gauthier et al. 2003; provided with each component analysis. braska north across southeastern Alberta, to Drum et al. 2015) may be important for a low of 8100 km2 for Chihuahuan Sandy addressing pervasive and critical threats to Plains Semi-Desert Grassland occurring in these grasslands, and to achieving desired TRENDS IN EXTENT OF MAJOR cross border region of Texas, New Mexico, conservation outcomes. GRASSLAND TYPES and adjacent Chihuahua and Coahuila, Mexico (Table 1). In this context, agencies forming the North Temperate grasslands of this trinational American Intergovernmental Commit- area include types ranging from prairies Long-term change in extent is estimated by tee on Cooperation for Wilderness and dominated by tall grasses and forbs, to comparing mapped current extent against Protected Areas Conservation (NAWPA) mixed-grass (tall and medium height these potential or historical estimates. Di- recently agreed to examine the adequacy grasses), to shortgrass prairies occurring rect conversion for agricultural use explains of the current North American conserva- in semi-arid climates of the southwestern much of the difference in areal extent for tion estate, considering key attributes of plains, and semi-desert grasslands found most types, with most extreme estimates ecological representation, connectivity, and throughout the Chihuahuan Desert. For this of long-term loss found in Texas Blackland resilience. By assessing achievements and effort, we utilized grassland classification Prairie (98%), Northern Tallgrass Prairie identifying gaps, efforts to restore and con- and map products based on the terrestrial (96%), Central Tallgrass prairies (92%), serve functionally resilient landscapes can ecological system classification of Nature- and Northern Fescue Mixed-grass Prai- be enhanced. This paper presents results Serve (Comer et al. 2003). This classifi- rie (87%). These are types occurring on of analyses focused on North America’s cation integrates floristic composition and productive soils in areas with significant temperate grasslands. geophysical settings to define units that summer rainfall and have, therefore, been formed the basis for regional and national intensively utilized for their agricultural Key elements to enhance landscape-scale land cover mapping in the United States productivity. Using just this one criteri- conservation include assessing resiliency, by federal interagency LANDFIRE (www. on—specific to historic long-term trends as defined by adequate representation of landfire.gov) and US Geological Survey’s in extent—from the International Union
198 Natural Areas Journal Volume 38 (2), 2018 Table 1. Long-term trends in extent and at-risk status of 12 major grassland types.
Major temperate grassland type Historical extent Current extent Percent loss to estimate (km2) estimate (km2) conversion Texas Blackland Tallgrass Prairie 41,400 670 98 Northern Tallgrass Prairie 157,200 6500 96 Central Tallgrass Prairie 242,000 20,100 92 Northern Great Plains Fescue Mixed-grass Prairie 137,000 18,000 87 Chihuahuan Sandy Plains Semi-Desert Grassland 8100 1600 80 Southeastern Great Plains Tallgrass Prairie 108,000 31,400 71 Central Mixed-grass Prairie 259,000 77,000 70 Western Great Plains Sand Prairie 107,300 38,000 65 Chihuahuan Loamy Plains Desert Grassland 38,300 14,400 62 Northwestern Great Plains Mixed-grass Prairie 620,900 307,500 50 Apacherian-Chihuahuan Semi-Desert Grassland and 249,400 152,200 39 Steppe Western Great Plains Shortgrass Prairie 259,000 188,000 27 Total 2,227,600 855,370 62 IUCN status based on long-term loss in extent Vulnerable Endangered Critically Endangered for Conservation of Nature (IUCN) Red some conservation concern in at least a (Zapus hudsonius preblei Krutzsch) [T2], List of Ecosystems (Keith et al. 2013), portion of the range) by Natural Heritage associated with prairie along riparian zones three types would be considered Critically field inventories in the United States and along the Rocky Mountain front. Fifty-nine Endangered, four types Endangered, and Canada were overlain on current extent tracked herptiles (both amphibians and rep- three types Vulnerable (Table 1). If assessed maps of the 12 major grassland types, and tiles) are associated with these grasslands, as one type of temperate grassland, a 62% the relative proportion of each species’ including the Colorado checkered whiptail estimated loss for these 12 types combined rangewide total was calculated. Table 2 (Aspidoscelis neotesselata Walker, Cordes, would qualify them as Vulnerable under summarizes numbers of species—grouped and Taylor) [G2], which is found in short- this IUCN criterion. as birds, mammals, herptiles, invertebrates, grass prairie and adjacent juniper savanna. and plants—tracked by NatureServe and Some 49 tracked invertebrates, primarily SPECIES OF CONCERN ASSOCIATED IUCN Conservation Status categories, consisting of insects such as the Dakota WITH THE MAJOR GRASSLAND and under legal protection in Canada, the skipper (Hesperia dacotae Skinner) [G2], TYPES United States, and Mexico that use the 12 are strongly associated with these prairie major grassland types. Eighty-three tracked types. Of over 500 tracked plant species Many plant and animal species that are de- bird species of concern are associated with associated with these grassland types, 124 pendent on these grassland types for core or these temperate grassland types, and 14 are categorized as G1–G3/T1–T3 range seasonal habitat are increasingly at risk and of these, such as mountain plover (Cha- at species or subspecies levels by Nature- identified on multiple lists of high-priority radrius montanus Townsend), are listed Serve. Some 16 are listed for protection in species for conservation attention. We used as threatened under Canadian protective Canada, and 13 are listed within the United site-based occurrence data from Natural status and associated with several of these States. An example is Mead’s milkweed Heritage Programs in the United States and major grassland types extending south into (Asclepias meadii Torr. ex Gray) [G2, Canada, and expert knowledge of habitat Mexico. Fifty-two tracked mammal species LT], a tallgrass prairie obligate, listed as associations in Mexican portions of the are associated with these grasslands, and Threatened in the United States. Chihuahuan Desert, to document species 12 of these, such as black-footed ferret of concern that are associated with one or (Mustela nigripes Audubon and Bach- Many at-risk species have been document- more of these 12 major grassland types. Be- man), are listed under protective status in ed for the Canadian prairies, which have cause these are upland grassland types, all Canada, the United States, and Mexico. been extensively converted to agriculture. species known to be limited to aquatic and Twelve mammal species fall into the G1T1 The vast area and diversity of circum- cave habitats were removed from analysis. (critically imperiled) to G3T3 (vulnerable) stances across the US grassland types are Over 150,000 documented occurrences of range of NatureServe Conservation status, indicated by varying numbers of at-risk 762 species tracked (i.e., considered of including Preble’s meadow jumping mouse species listed, and the relatively less in-
Volume 38 (2), 2018 Natural Areas Journal 199 tensively developed and less intensively studied grasslands of northern Mexico indicate fewer documented at-risk species. 6 6 8 6 26 NA
Espec.) GRASSLAND TYPE REPRESENTATION
(EN, T, Prot. IN CURRENT PROTECTED AREAS Mexico NOM-059 The IUCN has established a measure of conservation land status that includes six protected areas categories. These six cate- 6 6 gories range from Category I representing 52 16 10 14 “Strict Nature Reserve” to Category VI Candidate, EX = Extinct or Extirpated SC, C, EX) representing “Protected area with sustain-
COSEWIC (EN, T, able use of natural resources” (Dudley and Phillips 2006). However, many U.S. public
pecies levels. and tribal lands that are strongly regulated under environmental protection, endan- 4 3 4 1 25 13 gered species, and planning regulations are not included among IUCN categories, Candidate) although their management is strongly
US ESA (LE, LT, regulated under environmental protection, endangered species, and planning regula- tions. Ongoing discussions among NAWPA partners aim to fully harmonize land man- 3 3 8 8 8
30 agement designations on a trinational basis. VU, NT) Managed lands information (current in IUCN (CR, EN, 2016) applicable to this trinational study area were standardized to IUCN categories by the USGS Gap Analysis Program and provided for our analysis. Overlay of land 6 7 3 80 54 10 designations on current extent of the 12 atened, Protección Especial (Special Protection)
G3 or T3 major grassland types, and comparison of IC): EN = Endangered, T Threatened, SC Special Concern, C Vulnerable LT = Listed Threatened, Candidate for listing ed, VU = Vulnerable, NT Near Threatened these areas with estimates of both histor-
riled, imperiled, and vulnerable status at species (T) subs ical extent and current extent, resulted in calculations found in Table 3.
Investments in North American conserva- 2 1 1 65 49 11 tion lands are known to be concentrated
G2 or T2 among ecosystems with limited potential Imperiled for agriculture (Scott et al. 2001). These
major temperate grasslands are minimally y represented within designated conservation areas, with just 1.2% of the 2.2 million km2 historical extent estimate, or just 4 2 2 0 2 29 21 3.1% of the 855,370 km of estimated
Imperiled current extent, found within protected areas (Table 3). Major grassland types in the Chihuahuan Desert and surrounding region, including Apacherian-Chihuahuan Semi-Desert Grassland, appear to be most represented in designated protected areas, but the estimate of just 3% of historical extent for this type is still quite low as IUCN Red List status: CR = Critically Endangered, EN Endanger US Endangered Species Act (ESA) status: LE = Listed Endangered, Norma Official Mexicana 059 (NOM-059): EN = Endangered, T= Thre *NatureServe global status ranks G1–G3 indicate critically impe Committee on the Status of Endangered Wildlife in Canada (COSEW compared with other major vegetation types Total Plants Invertebrates Herptiles Mammals Birds Taxonomic group G1 or T1 Criticall Table 2. Numbers of at-risk species* associated with 12 major grassland types. grassland major 12 with associated species* at-risk of Numbers 2. Table in temperate North America (Aycrigg et al.
200 Natural Areas Journal Volume 38 (2), 2018 Table 3. Proportion of historical and current extent of each major grassland types represented in protected lands, as defined by IUCN Categories I-VI (as of 2016).
Major temperate grassland type Historical extent % in IUCN Current extent % in IUCN estimate (km2) Category I–VI estimate (km2) Category I–VI lands lands Apacherian-Chihuahuan Semi-Desert Grassland and 249,400 3 152,200 4.9 Steppe Chihuahuan Sandy Plains Semi-Desert Grassland 8100 3 1600 15 Northern Great Plains Fescue Mixed-grass Prairie 137,000 2 18,000 15 Northwestern Great Plains Mixed-grass Prairie 620,900 2 307,500 4 Chihuahuan Loamy Plains Desert Grassland 38,300 1 14,400 2.6 Western Great Plains Sand Prairie 107,300 1 38,000 2.8 Northern Tallgrass Prairie 157,200 0.4 6500 9.6 Western Great Plains Shortgrass Prairie 259,000 0.4 188,000 0.5 Southeastern Great Plains Tallgrass Prairie 108,000 0.3 31,400 1 Central Mixed-grass Prairie 259,000 0.1 77,000 0.3 Central Tallgrass Prairie 242,000 0.1 20,100 1.2 Texas Blackland Tallgrass Prairie 41,400 0.1 670 6 All Types Combined 2,227,600 1.2 855,370 3.1
2013). The estimated historical extent of or intact vegetation retains expected com- ing to characterize relative fragmentation eight in twelve of these major temperate position, structure, and dynamic process effects on the ecological condition of land- grassland types are represented by 1% characteristics of sites that have not been scapes (Riitters and Wickham 2003; Leu or less in designated protected areas, as altered by prior human land uses (Parrish et al. 2008; Theobald 2013). It integrates defined by IUCN Categories I-VI. Two et al. 2003). Substantial environmental data representing roads, land uses, and types that have experienced extremely degradation—due to wildfire suppression, mapped expressions of altered vegetation high proportional loss due to past land soil compaction, overgrazing, and climate to predict relative intactness in any given conversion, such as Northern Great Plains change—or disruption of biotic composi- area. This approach enables users to ex- Fescue Mixed-grass Prairie and Northern tion and processes, due to invasive species press assumptions at the site level about Tallgrass Prairie, appear to have relatively introductions and landscape fragmentation the relative ecological effect that each land high proportions of their current extent disrupting native species dispersal, have use type has (called a Site Impact Score) (15% and 9.6%, respectively) found within had substantial (albeit varying) impacts on and the potential effect as it diminishes protected areas. For contrast, many forest each grassland type (Samson et al. 2004). with distance from the site (called a Decay types occurring across the western states Score). Mapped information available for and northeast regions are represented in We developed two spatial models, first to percentages ranging from 5% up to over across the trinational area was compiled address ecological intactness of current 30%. Similar ranges are found among des- and standardized into 20 categories, orga- grassland areas, and second, to gauge rel- ert scrub vegetation types across the West. nized by (a) Transportation, (b) Urban and ative connectivity among major grassland Industrial Development, and (c) Managed patches. Practical measures of landscape in- and Modified Land Cover, each at 90-m LANDSCAPE INTACTNESS tactness often utilize remotely sensed data, pixel resolution. The resulting spatial index AND CONNECTIVITY AMONG which allow mapping landscape features combines the scores of all input layers and GRASSLANDS that result in habitat fragmentation and deg- their per-pixel values. Values close to 1.0 radation (Woolmer et al. 2008; Theobald imply relatively little ecological impact Land conversion and subsequent land et al. 2012). For example, fragmentation from surrounding land use. The result is protection tell only one part of the story of of natural habitat by human activities can a wall-to-wall grid surface of landscape trends in temperate grassland biodiversity. be modeled spatially, and used to identify condition values falling between 0.0 and A common consideration in landscape intact core areas and quantify the relative design for biodiversity conservation is to permeability of intervening areas. For this 1.0 (Figure 2). Independent data sets, document the relative condition or intact- effort, we used the NatureServe landscape including field observations of invasive ness, and key ecological processes such as condition model (Hak and Comer 2017). species and Natural Heritage Program field connectivity, in any given area in order to This model builds on the growing body of surveys, were used to calibrate and validate influence their selection for conservation published methods for ecological effects the model. See Hak and Comer (2017) for attention (Groves 2003). High condition assessment and spatial modeling, all aim- additional detail on the model.
Volume 38 (2), 2018 Natural Areas Journal 201 Figure 2. NatureServe landscape condition model applied to ecoregions currently supporting major grassland types, predicting a range of grassland quality.
202 Natural Areas Journal Volume 38 (2), 2018 By segmenting the landscape condition across a given landscape. These models to identify and prioritize places for con- model evenly across the 0.0–1.0 scale produced a per-pixel value for both in- servation across portions of this trinational range, and overlaying the spatial model tactness and relative connectivity across area have included state wildlife action on current grassland distributions, we the entire trinational study area. These plans (Mawdsley et al. 2016), and across summarized proportions of each grass- scores facilitate documentation of current political jurisdictions produced by The land type falling along the continuum of conditions, and were fed into subsequent Nature Conservancy (Groves 2003; Neely predicted landscape condition (Table 4). landscape conservation design to assist et al. 2006), the Nature Conservancy of Based on analyses reported in Hak and with selecting representative grassland Canada (Riley et al. 2007), Fish and Wild- Comer (2017), grassland areas scoring areas with highest potential for restoring life Joint Ventures (Giocomo et al. 2009), within the 0.61–1.0 range of landscape and maintaining ecological resilience. and Landscape Conservation Cooperatives condition are most likely to serve as “core (Olliff et al. 2016). areas” for grassland conservation. Among desert, semi-desert, and shortgrass types, a LANDSCAPE CONSERVATION New information and policy priorities relatively high proportion of the current ex- DESIGN among NAWPA agencies led to the desire tent occurs in better condition categories, as to revisit GPCA identification in this re- compared with other types. As anticipated, Grassland Priority Conservation Areas gion. First, much new mapped information tallgrass prairie types tended to score most (GPCAs) define areas where conserva- pertaining to grassland types and landscape poorly over greater proportions of their tion attention could be concentrated to conditions became available, enabling a current extent than other types. Tallgrass advance biodiversity conservation goals. focus of grassland diversity to complement types occur in areas with lowest landscape Their identification does not imply any prior efforts emphasizing grassland bird condition categories (0.40–0.00), and will particular conservation action, or suggest species. Second, there was a common tend to be most challenging for ecological change in ownership. However, the lands desire to enable national reporting of restoration. within identified GPCAs present conserva- progress toward “Aichi Targets” under the tion opportunities to be considered, along United Nations Convention on Biodiversity The second of the two models aimed with current and established management (CBD). Nations who signed on to the Con- to measure relative connectivity among priorities, to strive for compatibility with vention met in Nagoya, Aichi Prefecture, grassland patches. We used the NatureServe biodiversity values they support today, or Japan, in 2010 and committed to a series of landscape condition model as a “resistance” could feasibly support through ecological national conservation targets. Aichi Target surface to model relative connectivity restoration. Grassland Priority Conserva- 11 states that “By 2020, at least 17 percent among grassland patches using methods tion Areas have been previously identified of terrestrial and inland water areas… and tools designed by Theobald et al. across this trinational area (PCAP 1998; especially areas of particular importance (2012). That is, with increasing landscape Gauthier et al. 2003; Pool and Panjabi for biodiversity and ecosystem services, are fragmentation, there is increasing “resis- 2011), each with a primary focus on con- conserved through effectively and equita- tance” to lateral movement by species serving grassland species. Similar efforts bly managed, ecologically representative
Table 4. Proportion of current extent of each major grassland types predicted to occur across a range of quality and condition, based on the NatureServe landscape condition model.
Grassland type 0.81–1.0 0.61–0.80 0.41–0.60 0.21–0.40 0.00–0.20 (best (moderate– (moderate ) (moderate– (worst condition ) good ) poor ) condition ) Chihuahuan Sandy Plains Semi-Desert Grassland 72% 19% 6% 2% 1% Chihuahuan Loamy Plains Desert Grassland 61% 24% 8% 5% 3% Apacherian-Chihuahuan Semi-Desert Grassland and Steppe 57% 23% 13% 5% 2% Western Great Plains Sand Prairie 51% 15% 16% 13% 6% Western Great Plains Shortgrass Prairie 45% 22% 15% 13% 6% Northwestern Great Plains Mixed-grass Prairie 36% 23% 24% 12% 5% Northern Great Plains Fescue Mixed-grass Prairie 28% 20% 40% 11% 2% Southeastern Great Plains Tallgrass Prairie 18% 13% 23% 28% 19% Central Mixed-grass Prairie 10% 17% 28% 29% 16% Texas Blackland Tallgrass Prairie 6% 10% 18% 30% 36% Central Tallgrass Prairie 4% 7% 28% 34% 28% Northern Tallgrass Prairie 3% 9% 28% 35% 25% All Types Combined 39% 20% 20% 14% 7%
Volume 38 (2), 2018 Natural Areas Journal 203 and well-connected systems of protected contribution toward grassland conserva- with extensive rock outcrops in its core areas and other effective area-based con- tion goals. That is, hexagons with high distribution of the Flint Hills has limited servation measures, and integrated into areal grassland extent and high landscape intensive conversion of this type for agri- the wider landscape and seascape.” The intactness scored highest for initial site culture. Occurring in relatively small and specific forms of defining “17 percent” selection. Highest-scoring hexagons were fragmented areas, the Chihuahuan Sandy and “ecologically representative” area is selected preferentially in sufficient num- Plains Semi-Desert Grassland was not rep- left to each reporting country. bers to initially advance toward the 17% resented at or above the 17% milestone. As representation goal. Grassland experts from anticipated, it was also quite challenging to We used mapped information on each each jurisdiction reviewed and refined site adequately represent other tallgrass prairie major grassland type to assess relative selection. Existing managed areas, land use types occurring across the more humid, representation in conservation invest- intensity and connectivity maps, and at-risk eastern portions of the region, as these ments under Aichi Target 11. We also species distributions, and already identified types have been extensively converted for utilized estimates of historical extent of priority conservation areas, were all used agriculture (Table 1). Identified GPCAs each grassland type to assist with gauging as ancillary information to refine areas se- encompass Central Tallgrass Prairie (4%), progress relative to the Aichi Target 17% lection aiming to efficiently represent each Northern Great Plains Fescue Mixed-grass percent-area milestone. Grassland Priority grassland type in landscapes of greatest Prairie (4%), Northern Tallgrass Prairie Conservation Areas (GPCAs) were iden- complementary conservation value. Once (2%), and Texas Blackland Tallgrass Prairie tified to build upon existing conservation selected, hexagons were finalized, and adja- (1%). Undoubtedly, these types bring the lands, and to use Aichi Target 11 as one cent clusters were systematically identified greatest challenges for restoration and practical milestone for temperate grassland and labeled to form 177 distinct GPCAs. representation of habitat conditions for the representation. This approach in landscape high diversity of species they each support. conservation design helped to reflect the Results include representative areas for Taken together, approximately 15% of diversity of major grassland-dominated all 12 types, and achieve levels of rep- estimated historical extent of all 12 major landscapes and habitat requirements for resentation that meet, or surpass, Aichi temperate grassland types occurs in newly many associated species. Target 11 (Figure 3). Others that do not identified GPCAs (Table 5). While just achieve those levels of representation below the 17% Aichi Target 11 milestone, Two major site selection strategies were are included in landscapes with at least the selection of GPCAs still represents all used: some substantial potential for successful major grassland types in areas most likely 1) Efficient type representation strategy restoration. Table 5 includes a high-level to support successful conservation. for grassland types that retain large pro- summary of grassland type representation portional area and/or relatively large intact within the newly identified GPCAs. The Tables 6–8 provide a high-level summary blocks based on current grassland condition Nebraska Sand Hills includes an extensive of conditions associated with the 177 and connectivity. and contiguous area of Western Great identified GPCAs, again organized by the 2) Emphasis on consolidation of high-val- Plains Sand Prairie, so it is feasible to predominant grassland type they contain. ue sites for restoration of types that have represent the highest proportion of this Table 6 summarizes spatial information experienced extreme loss through land type. Our results indicate that about 27% used in the landscape condition model. conversion since 1800. of estimated historical extent of this type The table is sorted based on the estimated occurs within identified GPCAs. Six other total area of each grassland type occurring Landscape design was completed by using major grassland types met or surpassed the within identified GPCAs. While purposely a grid of nearly 48,000 hexagonal spatial 17% milestone of Aichi Target 11. These designed to avoid and minimize overlap analysis units, each 100 km2 in size. Given include Northwestern Great Plains Mixed- with current intensive—and potentially the trinational scale of analysis, areas iden- grass Prairie (21%), Chihuahuan Loamy threatening—land uses, each GPCA will tified were no smaller than 100 km2, with Plains Desert Grassland (19%), Central inevitably include a footprint of some the presumption that subsequent planning Mixed-grass Prairie (18%), Southeastern level of current land use. Knowledge of decisions would bring local knowledge to Great Plains Tallgrass Prairie (18%), West- these patterns can be helpful for clarifying specify areas for conservation action. We ern Great Plains Shortgrass Prairie (17%), conservation strategies. Patterns in the applied common systematic conservation and Apacherian-Chihuahuan Semi-Desert occurrence and abundance of different planning methods (Groves and Game 2016) Grassland (17%). These types dominate road types, development intensity, wells and spatial optimization algorithms (Ball et the relatively dry-to-arid portions of the and mines, current croplands, pasture and al. 2009) augmented by expert review and study area across the western Great Plains ruderal vegetation, and major centers of refinement. In the first and primary step, for and extending throughout the Chihuahuan invasive plant species are summarized. On each hexagon, a measure of grassland ex- Desert. Only the Southeastern Great Plains average, the GPCAs avoid primary roads tent (standardized to 0.0–1.0) was divided Tallgrass Prairie represents tallgrass prairie and highways, but all include varying levels by the inverse of the landscape intactness types in this group. It occurs in relative- of more local and unpaved road networks. score (also on the scale of 0.0–1.0) to ly humid areas of eastern Kansas and Current cropland makes up a substantial provide a relative index of their potential Oklahoma, but relatively shallow soils component of GPCAs that represent mixed-
204 Natural Areas Journal Volume 38 (2), 2018 Figure 3. Grassland Potential Conservation Areas (GPCAs) based on representation of 12 major temperate grassland types (ordered north to south).
Volume 38 (2), 2018 Natural Areas Journal 205 Table 5. Proportion of historical extent of each major grassland types represented in identified Grassland Potential Conservation Area (GPCA).
Grassland type Historical Current Number of Grassland % Historical extent extent GPCAs km2 within extent in estimate estimate GPCAs GPCAs (km2) (km2) Western Great Plains Sand Prairie 107,300 38,000 7 29,200 27 Northwestern Great Plains Mixed-grass Prairie 620,900 307,500 26 128,400 21 Chihuahuan Loamy Plains Desert Grassland 38,300 14,400 10 7300 19 Central Mixed-grass Prairie 259,000 77,000 9 39,200 18 Southeastern Great Plains Tallgrass Prairie 108,000 31,400 5 19,300 18 Western Great Plains Shortgrass Prairie 259,000 188,000 14 50,700 17 Apacherian-Chihuahuan Semi-Desert Grassland and 249,400 152,200 34 42,000 17 Chihuahuan Sandy Plains Semi-Desert Grassland 8100 1600 16 1000 12 Central Tallgrass Prairie 242,000 20,100 8 10,000 4 Northern Great Plains Fescue Mixed-grass Prairie 137,000 18,000 29 5400 4 Northern Tallgrass Prairie 157,200 6500 15 3500 2 Texas Blackland Tallgrass Prairie 41,400 670 4 400 1 Total 2,227,600 855,370 177 342,400 15% grass prairie types, as these large GPCAs compass over 46,500 km2 and could be was intended as a relatively rapid, subcon- tend to occur in landscapes supporting reviewed within each GPCA to determine tinental-scale analysis, and so there are both cropland and pasture. This mixture of where additional levels of biodiversity several key considerations for building on current land uses is exemplified by North- conservation could be emphasized. Con- these results to advance conservation and western Great Plains Mixed-grass Prairie. servation easements represent the transfer build resiliency. GPCAs representing this type extend across of development rights for the benefit of the northern Great Plains where the mix conservation values. While relatively small, First, since we focused primarily on major of land use is well supported. these areas are currently most concentrated grassland types, there are many less abun- in Northwestern Great Plains Mixed-grass dant ecosystem types not yet treated. We Table 7 summarizes existing protected Prairie, Northern Tallgrass Prairie, and anticipate that most characteristic upland areas as they contribute to the GPCAs. Western Great Plains Shortgrass Prairie, ecosystems, as well as needs for most grass- Following from Table 3, a relatively small respectively. If one combines these current land birds, have been represented within proportional area of the newly identified conservation investments, they amount the GPCAs. However, much complemen- GPCAs are currently protected under one of to over 52,000 km2, or about 15% of the tary analysis and priority-setting in this the I–VI categories from IUCN, with 1.2%, newly identified GPCA area. region has centered on wetland ecosystems or a total of 40,496 km2, occurring in lands and the diversity of species they support designated into at least one of the IUCN The priority areas already identified by (e.g., Beyersbergen et al. 2004). Similarly, status categories. Of these, IUCN category The Nature Conservancy and the Nature systematic prioritization, assessment, and VI (Protected area with sustainable use of Conservancy of Canada include substantial monitoring of lake and stream ecosystems natural resources) is most common with lands and encompass fully 84% of the land (e.g., Stagliano 2006) could provide anoth- 2 23,133 km , or 54% of the total GPCA area in newly identified GPCAs. This result er important complement to this analysis. lands, designated as such. was to be expected in that the purpose and Second, this effort did not directly address intent of those planning processes overlaps Table 8 summarizes information on com- the rates of landscape change or use fore- substantially with the goals of this effort, plementary conservation investments and casting to gauge the risk of future land use and their mapped boundaries were taken priorities related to the newly identified GP- patterns. While land use change tends to into account as sites were identified. CAs associated with each major grassland follow existing patterns, and those patterns type. “Other Conservation Areas” tend to were used directly in GPCA selection, include US public lands that have a clear BUILDING ON THIS ANALYSIS there can be land use trends that are dif- multiple-use mandate, but do not fall within ficult to foresee, such as trends in energy any of the IUCN protected land categories. While this process followed common development stimulated by changes in These areas, such as National Grasslands methods and best practices used in region- technology or policy. For example, con- managed by the US Forest Service, en- al-scale landscape conservation design, it centrated renewable energy development
206 Natural Areas Journal Volume 38 (2), 2018 can rapidly expand due to advances in technology and placement of transmission ) 2
– infrastructure (Lewis and Wiser 2007). 1
). Changes in commodity prices have also 2 (km plants
Invasive brought increasing unpredictability in patterns of land conversion in this region 0 – ) 19 720
2 (Wright and Wimberly 2013). Furthermore, we have not specifically addressed potential (km ruderal near-term direct effects of climate change Pasture & on grassland biodiversity or interactions of changing climate with other ecosystem 3 ) 2 99 4932 – 45 343stressors – such as displacement of native
(km species by invasive species expansion. cropland
Agriculture Following emerging recommendations for best practices (Gillson et al. 2013; Comer 2018), additional steps could be applied to ) 46 19 22,711 4767 7 47 79 2 evaluate and potentially modify identified
(km GPCAs to maximize their resilience in the (High–Low)
Development face of climate change.
Third, we were limited to available spatial data and modeling to gauge relative intact-
Unpaved ness and connectivity among grasslands. roads (km) There are undoubtedly land use legacies and current impacts not adequately reflect- ed in spatial data (Gauthier et al. 2003).
(km) For example, grazing pressure on these
Local roads grasslands has varied substantially over decades, and its effects may not always correlate well with current patterns in land use and infrastructure. Therefore, our mea- sures of landscape intactness could mask Secondary roads (km) true conditions on the ground where past land uses have severely degraded grassland condition in areas that are now extensive, contiguous, and remote. Regional-scale
Primary remote sensing indices (e.g., Muldavin et roads (km) al. 2001) may prove valuable in further
) assessing landscape conditions where 2 400 480 2331 22,847 84 938 1467 5037 1 3500 709 4693 39,188 3355 1700 16,167 3146 489 5400 2617 37,212 588 37,212 1000 352 364 6503 1317 7300 – 703 10,260impacts 746 are more diffuse. 10,000 234 3215 27,909 605 1059 6731 3254 39,000 881 7218 90,084 2883 2369 13,757 2273 6156 19,300 1127 6277 56,872 1203 2342 3342 3898 – 50,70042,000 200 656 2277 33,571 710 11,062 3821 11,321 598 3785 191 1784 km ( 128,400 3650 59,365 79,285 67,232 1652 46,698 25,908 3915 342,400 10,972 127,310 404,159 134,021 11,363 116,792 54,068 13,07 area in GPCAs Grassland All three of these areas should be con- sidered and appropriate actions should be taken to make the most of our analysis. These results should be suitable for use by the NAWPA agencies as they engage with
rass Prairie partners and stakeholders in conservation g planning, protected area expansion, habitat
rass Prairie restoration, and environmental monitoring. g rass Prairie g rass Prairie Specific conservation objectives for each g GPCA may then be collaboratively devel- oped based on grassland type, associated species, threats, relative cost of conserva- orthern Tall rass Prairie Total Chihuahuan Sandy Plains Semi- Desert Grassland Texas Blackland Tall Grassland type Western Great Plains Sand PrairieSoutheastern Great Plains Tallgrass Prairie Central Tall 29,200Chihuahuan Loamy Plains Desert Grassland Northern Great Plains Fescue Mixed- 66g N 2946 25,990 4242 514 1988 3 Northwestern Great Plains Mixed- grass Prairie Western Great Plains Shortgrass Prairie Apacherian-Chihuahuan Semi-Desert Grassland and Steppe Central Mixed- tion action, associated stakeholder needs,
Table 6. Summary information on land uses and threats within GPCAs associated with each major grassland types (linear features expressed as km distance, areal features as km as features areal distance, km as expressed features (linear types grassland major each with associated GPCAs within threats and uses land on information Summary 6. Table and other factors.
Volume 38 (2), 2018 Natural Areas Journal 207 CONCLUSIONS
We have demonstrated a systematic ap-
35 proach to use major grassland ecosystem 896 550 245 422
1048 1563 type distributions as a novel approach Combined IUCN I-VI for advancing toward, and beyond, Aichi Target 11 for ecosystem representation. Grassland Potential Conservation Areas – – – – – – – can serve as one important focus for 816 2696 318 2001 070 7317 5 5317 5913 collaborative conservation action to build resources IUCN VI landscape resiliency in the face of accel- use of natural Protected area with sustainable erating environmental change.
Much of the Great Plains and Chihuahuan 30 ). 2 122 525 489 632 375 245 470 290 Desert is dominated by private lands. Con- IUCN V Protected servation strategies throughout this region Landscape are inherently collaborative, often involv- ing partnerships and programs aiming to provide technical assistance and financial – 5 48 45 12 1487 incentives that encourage natural resource 730 235 1375 229 405 130 Area 1219 5145 8736 23,133 42,496 2089 2696 11,612 19,302 conservation. The emergence of ecosystem IUCN IV
Management service markets across North America Habitat/Species may provide additional opportunities to advance conservation within and among GPCAs where markets for soil carbon,
4 – 1 – – – water resources, and wetland or grassland 13 73 Natural Feature
IUCN III habitat mitigation generate revenue for land Monument or owners and managers. –– – – – – – – – – From this analysis, we have produced an 817 25 166 10
Park online conservation atlas to share project National IUCN II data and provide input to subsequent agen- cy and partner decision making processes (www.DataBasin.org). The site includes a series of maps, text, and tabular summa- 1 0 – – – 31 64 ries to describe the GPCAs in terms of 173 512 2801 2520 161 IUCN I the grassland types, existing conservation Reserve or Wilderness Strict Nature investments, and threats associated with their conservation. It may be accessed by the public and partners interested in biodi- versity conservation across this trinational area. Supplementary materials associated with this paper include descriptions of each rass Prairie 1332 1537 36 g major grassland type, tabular summaries of rass Prairie
g grassland associated species, and GPCA rass Prairie
g site summaries, and are found on the DataBasin site. rass Prairie g Plains Desert Grassland y ACKNOWLEDGMENTS rass Prairie g rass Prairie g rass Prairie g We wish to acknowledge the support from e
pp NAWPA members and facilitators, espe- cially US Fish and Wildlife Service staff
orthern Tall orthwestern Great Plains Mixed- John Schmerfeld and Nancy Roeper, and Apacherian-Chihuahuan Semi-Desert Grassland and Ste N N Western Great Plains Short Total Northern Great Plains Fescue Mixed-grass Prairie 64 Central Tall Chihuahuan Loam Chihuahuan Sandy Plains Semi-Desert Grassland 625 Texas Blackland Tall Western Great Plains Sand Prairie Central Mixed- Grassland type Southeastern Great Plains Tall Adam Hanson from the Wild Foundation. Table 7. Summary information on current conservation investments within GPCAs associated with each major grassland types (km types grassland major each with associated GPCAs within investments conservation current on information Summary 7. Table USGS and Gap Analysis Program staff
208 Natural Areas Journal Volume 38 (2), 2018 Table 8. Summary information on Other Conservation Areas, conservation easements, and TNC or NCC conservation priorities in GPCAs associated with each major grassland types (km2).
Grassland type Other Conservation TNC or NCC conservation area easement (km2) priority (km2) (km2) Apacherian-Chihuahuan Semi-Desert Grassland 5150 143 43,997 and Steppe Central Mixed-grass Prairie 319 42 39,155 Central Tallgrass Prairie 34 224 7711 Chihuahuan Loamy Plains Desert Grassland 2065 244 7509 Chihuahuan Sandy Plains Semi-Desert Grassland 2472 534 7525 Northern Great Plains Fescue Mixed-grass Prairie – – 891 Northern Tallgrass Prairie 700 1166 7099 Northwestern Great Plains Mixed-grass Prairie 26,485 2145 68,031 Southeastern Great Plains Tallgrass Prairie 509 207 22,971 Texas Blackland Tallgrass Prairie 79 16 1781 Western Great Plains Sand Prairie 1484 252 53,411 Western Great Plains Shortgrass Prairie 7250 1002 35,727 Total 46,547 5975 295,807
Kevin Gergely, Leah Dunn, Steve Williams, Esteban Muldavin has been with Natural Patrick Comer is Chief Ecologist for and Lisa Johnson provided protected area Heritage NM since 1991 and as the Senior NatureServe. Pat helped pioneer the use Ecologist and the Director since 2007. He maps used in this analysis. Healy Hamil- of standardized terrestrial ecosystem is an expert in arid-lands ecological as- ton (NatureServe) provided programmatic classifications for national mapping in sessment, vegetation dynamics, and climate leadership and supported partner collab- the United States and continental-scale change modelling, and the maintenance of oration throughout this effort. We also efforts across the Americas. Since the a statewide database for holding vegetation wish to acknowledge the expert review mid-1990s, Pat has been advancing meth- classification and monitoring data. and input from the following specialists ods for ecoregion-scale assessment and from NAWPA agencies and NatureServe conservation design, developing methods Jason Singhurst has conducted field re- network from throughout the project area. applied by The Nature Conservancy, and search on rare plant communities and rare These included Christian Artuso (Bird for US federal agencies such as the Bureau plants across Texas as a Botanist/Ecologist Studies Canada), Andrea Cantú Garza of Land Management. for the past 22 years for Texas Parks and (Pronatura Noreste), Pat Fargey (Alberta Wildlife Department. In 2008, he co-au- Environment and Parks), Craig Freeman Jon Hak is Senior Ecologist/GIS Analyst thored a book on Rare Plants of Texas. (Kansas Natural Heritage Program), Anne for NatureServe. He develops and applies Halford (Bureau of Land Management), analytical tools for species and ecosystem Tyler Johnson (US Forest Service), Ignacio mapping, landscape condition and connec- March (CONANP), Mauricio De la Maza tivity modeling, planning, and monitoring LITERATURE CITED (Pronatura Noreste), Richard Pither (Parks for biodiversity conservation. Aycrigg, J.L., A. Davidson, L.K. Svancara, K.J. Canada), Rick Schneider (Nebraska Nat- Gergely, A. McKerrow, and J.M. Scott. 2013. ural Heritage Program), Gerry Steinauer Kelly Kindscher is a Professor in Envi- Representation of ecological systems within (Nebraska Natural Heritage Program), ronmental Studies and Senior Scientist at the protected areas network of the continen- Linda Vance (Montana Natural Heritage the Kansas Biological Survey, both at the tal United States. PLoS One 8(1):e54689. Program), Wendy Velman (Bureau of Land University of Kansas. His work has focused Ball, I.R., H.P. Possingham, and M. Watts. 2009. Management), Neal Niemuth (US Fish and on prairie and wetland identification, as- Marxan and relatives: Software for spatial conservation prioritisation. Pp.185-195 in A. Wildlife Service), and Cliff Wallis (Alberta, sessment, and restoration, and he is well Moilanen, K.A. Wilson, and H. Possingham, Canada). Mary Harkness and Mary Russo known for his work as an ethnobotanist eds., Spatial Conservation Prioritisation: (NatureServe) provided database support to and publications on edible and medicinal Quantitative Methods and Computational produce summary tables for this analysis. plants of the Great Plains and Midwest. Tools. Oxford University Press, Oxford, UK.
Volume 38 (2), 2018 Natural Areas Journal 209 Beyersbergen, G.W., N.D. Niemuth, and M.R. Flight Conference: Tundra to Tropics, 13-16 G. Belew, F. Pusateri, et al. 2006. Central Norton, coords. 2004. Northern Prairie & February 2008, McAllen, TX. [International Shortgrass Prairie Ecoregional Assessment Parkland Waterbird Conservation Plan. A Conferences of] Partners in Flight (PIF). and Partnership Initiative. The Nature Con- Plan Associated with the Waterbird Con-
210 Natural Areas Journal Volume 38 (2), 2018 America’s biological diversity? Ecological Theobald, D.M., S.E. Reed, K. Fields, and M. M. Anderson, R. Baldwin, A. Morgan, and Applications 11:999-1007. Soulé. 2012. Connecting natural landscapes E. Sanderson. 2008. Rescaling the human Stagliano, D. 2006. Freshwater Measures for using a landscape permeability model to pri- footprint: A tool for conservation planning at the Northern Great Plains Steppe Ecoregion oritize conservation activities in the United an ecoregional scale. Landscape and Urban of Montana. Montana Natural Heritage States. Conservation Letters 5:123-133. Planning 87:42-53. Program, Helena. With, K.A., A.W. King, and W.E. Jensen. 2008. Wright, C.K., and M.C. Wimberly. 2013. Recent Theobald, D. 2013. A general model for quan- Remaining large grasslands may not be suf- land use change in the Western Corn Belt tifying ecological integrity for landscape ficient to prevent grassland bird declines. Bi- threatens grasslands and wetlands. Proceed- assessments and U.S. application. Landscape ological Conservation 141(12):3152-3167. ings of the National Academy of Sciences Ecology 28:1859-1874. Woolmer, G., S.C. Trombulak, J. Ray, P. Doran, 110(10):4134-4139.
Volume 38 (2), 2018 Natural Areas Journal 211