Population and Habitat Objectives for Avian Conservation in California’s Central Valley Grassland–Oak Savannah Ecosystems

In California’s Central Valley, grassland and oak savannah ecosystems provide multiple economic and social benefits, ecosystem services, and vital bird habitat. There is a growing interest in protecting, restoring, and managing these ecosystems, and the Central Valley Joint Venture (CVJV) provides leadership in the formulation of conservation goals and objectives. We defined a long-term goal of protecting, restoring, and managing Central Valley grassland and oak savannah ecosystems so that they are capable of supporting genetically robust, self-sustaining, and resilient wildlife populations. To measure progress toward this goal, we selected a suite of 12 landbird focal species that primarily breed in grasslands and oak savannahs as indicators of the state of these ecosystems on the Central Valley floor (primary focus area) and in the Central Valley’s surrounding foothills (secondary focus area). Using data on current densities and habitat extent, we estimated that at least three of the focal species populations in the primary focus area and at least two of the focal species populations in the secondary focus area are currently small (<10,000 individuals) and may be vulnerable to extirpation. Furthermore, at least two species appear to have steeply declining population trends. We defined long-term (100-year) population objectives for each focal species that we expect to meet the goal of genetically robust, self-sustaining, and resilient populations. We then estimated corresponding short-term (10-year) habitat objectives of 4,183 ha of additional grassland and 3,433 ha of additional oak savannah that will be required to make progress toward the long-term objectives. We expect that habitat restoration and enhancement efforts aimed at reaching these long-term conservation objectives will result in improvements to the function of Central Valley grassland and oak savannah ecosystems.

the Central Valley Joint Venture (CVJV) provides leadership in the formulation of conservation goals and objectives. We defined a long-term goal of protecting, restoring, and managing Central Valley grassland and oak savannah ecosystems so that they are capable of supporting genetically robust, self-sustaining, and resilient wildlife populations. To measure progress toward this goal, we selected a suite of 12 landbird focal species that primarily breed in grasslands and oak savannahs as indicators of the state of these ecosystems on the Central Valley floor (primary focus area) and in the Central Valley's surrounding foothills (secondary focus area). Using data on current densities and habitat extent, we estimated that at least three of the focal species populations in the primary focus area and at least two of the focal species populations in the secondary focus area are currently small (<10,000 individuals) and may be vulnerable to extirpation. Furthermore, at least two species appear to have steeply declining population trends. We defined long-term (100-year) population objectives for each focal species that we expect to meet the goal of genetically robust, self-sustaining, and resilient populations. We then estimated corresponding short-term (10-year) habitat objectives of 4,183 ha of additional grassland and 3,433 ha of additional oak savannah that will be required to make progress toward the long-term objectives. We expect that habitat restoration and enhancement efforts aimed at reaching these long-term conservation objectives will result in improvements to the function of Central Valley grassland and oak savannah ecosystems.

INTRODUCTION
Grassland and oak savannah ecosystems are an important component of California's Central Valley uplands, particularly the open country "bathtub 2 VOLUME 15, ISSUE 1, ARTICLE 6 ring" of low-elevation foothills and rangelands that surround the perimeter of the valley floor ( Figure 1). In addition to providing important landbird habitat (Verner 1980;CPIF 2000), these ecosystems provide a number of functions, including nutrient and water cycling, pollinator population support, and food and fiber production through livestock operations (Havstad et al. 2007;Kroeger et al. 2009;Chaplin-Kramer et al. 2011). However, roughly 60% of the Central Valley's historic grasslands have been lost as a result of conversion to intensive agriculture (e.g. orchards, vineyards, and row crops) and urban development (CPIF 2000;DGP-GIC 2003). Similar historical data on the extent of oak savannah ecosystems in the Central Valley are lacking, but the magnitude of loss is probably similar. Today, both grassland and oak savannah ecosystems are still at risk of conversion to land uses that may not provide the suite of ecosystem services they currently generate (Cameron et al. 2014;Byrd et al. 2015). Furthermore, rapid climate change threatens to degrade a number of ecosystem services, including forage production and carbon sequestration (Kueppers et al. 2005;Shaw et al. 2011;Byrd et al. 2015). Given these threats, and the multiple benefits of conserving these ecosystems, there is a strong and growing interest in protecting, restoring, and managing grasslands and oak savannahs in California. For these efforts to be successful, clear goals and objectives are necessary.
The Central Valley Joint Venture (CVJV; http://www. centralvalleyjointventure.org/), established in 1988, is a coalition of 20 state, federal, and private partners with the common goal of providing sufficient habitat for migrating and resident birds in the Central Valley of California. Setting explicit conservation objectives has become a standard approach to conservation planning and implementation, helping to unify stake-holders (Williams and Madsen 2013), make conservation more efficient (Carwardine et al. 2009), prioritize investments (Margules and Pressey 2000), and demonstrate success, which inspires support from the public and funders. Increasing the population size of a particular species or group of focal species is a common conservation objective, either for recovering imperiled species or, in the case of focal species, as an indicator of habitat condition (Lambeck 1997;Carignan and Villard 2002;Chase and Geupel 2005;Sanderson 2006). The underlying concept of the focal species approach is simply that by managing habitat to maintain the conditions that support populations of focal species, the appropriate conditions will be maintained to support diverse and healthy ecosystems (Chase and Geupel 2005). The CVJV has adopted this approach, defining population objectives for waterfowl, shorebirds, waterbirds, and riparian landbirds as indicators of the condition of Central Valley's wetlands and riparian habitat (CVJV 2006). While it is unlikely that monitoring any single taxon can capture all dimensions of ecosystem function (Lindenmayer et al. 2015), birds are recognized as useful ecological indicators because many species are easily and inexpensively detected using standardized sampling; these species reflect a wide variety of habitat conditions; and accounting for and maintaining many species with different ecological requirements can be used to implement landscape-scale conservation strategies and maintain the ecosystem services that birds provide (Hutto 1998;Alexander et al. 2007;Whelan et al. 2008;Ortega-Alvarez and Lindig-Cisneros 2012).
Here, we describe our process for setting longterm (100-year) and short-term (10-year) CVJV conservation goals and objectives for grassland and oak savannah ecosystems. We defined a long-term goal of protecting, restoring, and managing Central Valley grassland and oak savannah ecosystems so that they are capable of supporting genetically robust, self-sustaining, and resilient wildlife populations (Sanderson 2006;Redford et al. 2011;Dybala et al. 2017, this volume). To define the specific, measurable conservation objectives required to meet this long-term goal, we first estimated the current extent of Central Valley grassland and oak savannah vegetation, and the current population size and density of 12 focal species. Then we defined long-term population objectives for each species, and the long-term density and habitat objectives required to achieve the population objectives. We also generated short-term (10-year) density and habitat objectives as milestones for tracking progress.
By setting these objectives, and guiding resource management plans toward achieving these objectives, we expect to improve Central Valley grassland and oak savannah ecosystems overall, thus benefitting species of conservation concern, helping keep  4 VOLUME 15, ISSUE 1, ARTICLE 6 common birds common, and benefitting grassland and oak savannah wildlife communities beyond birds.

Study Area
The Central Valley of California is a large valley bounded by the Sierra Nevada mountain range to the east and the Coast Ranges to the west (Figure 1). The valley is sub-divided into the northern Sacramento Valley and southern San Joaquin Valley, each drained by a major river of the same name. The confluence of these two rivers forms the Sacramento-San Joaquin River Delta. The Tulare Basin, at the southern end of the Central Valley, is a terminal basin distinct and separate from the rest of the San Joaquin Valley, and its rivers once drained into Tulare Lake and several smaller lakes and sloughs. Except in significant flood years, the Tulare Basin is now mostly dry as a result of water diversion and conversion to agriculture. The geographic scope of the CVJV extends across the full extent of the valley floor, and is divided into two focus areas for the purposes of grassland and oak savannah conservation planning (Figure 1). The primary focus area includes the entire floor of the Central Valley and the Carrizo Plain, and is largely delineated by the Jepson boundary for the Great Central Valley region (Hickman 1993). The secondary focus area encompasses the foothills that surround the valley floor, and generally extends to the crests of surrounding watersheds.

Focal Species
We selected 12 focal species representative of grassland and/or oak savannah ecosystems, five of which principally use grassland vegetation as their preferred habitat and seven of which principally use oak savannah vegetation as their preferred habitat (Table 1). These species were selected because they: (1) use valley grassland or oak savannah as breeding habitat in the Central Valley, (2) warrant special management status, or have experienced population declines or reductions in breeding range in the Central Valley, or (3) are useful for monitoring the effects of management actions in Central Valley grassland or oak savannah ecosystems, in part because they are common enough to provide sufficient sample sizes for analyses. These focal species also represent a range of life histories and vegetation associations that reflect a continuum of habitat needs (Table 1), and their study provides information about different habitat attributes of Central Valley grassland and oak savannah ecosystems.

Grassland and Oak Savannah Habitat
Each of the focus areas contains both grassland and oak savannah vegetation types. Thus, estimating the current extent of each vegetation type in each focus area required both a vegetation data source with broad coverage of California and a definition of grasslands and oak savannahs. We used a recently updated vegetation layer compiled from many sources to provide the "best available" current California land cover information in one layer (CAL-FIRE 2015, unpublished GIS data, see "Notes"). In this layer, land cover is classified according to the California Wildlife Habitat Relationships classification system, including estimates of tree canopy closure for woodland or forest classes: sparse (10% to 24%), open (25% to 39%), moderate (40% to 59%), and dense (60% to 100%).
There is no single unifying and agreed upon definition of "grassland" or "oak savannah," though Stromberg et al. (2007) describes grasslands as dominated by grasses and other herbaceous plant species, and savannahs as a "grassland containing widely spaced trees." Our definition of "grassland" is a landscape dominated by grasses and other herbaceous plant species with <10% woody cover, and "oak savannah" as woodlands with sparse (10% to 40%) tree cover, where the dominant tree species are oaks (Quercus spp.) and primarily grassdominated understories. Since alpine and mountain meadow grasslands are ecologically distinct from valley grasslands, we restricted our estimation of grassland and oak savannah habitat to below the 3,000-ft elevation mark.
We reclassified the vegetation layer to include annual grasslands (AGS), perennial grasslands (PGS), and pasture (PAS) as grasslands; and sparse

Species Densities, Population Sizes, and Trends
For eight of the 12 focal species, we estimated breeding densities in grassland and oak savannah vegetation types in each focus area by analyzing point count data collected from 2002-2015 by Point Blue Conservation Science and partners from 511 points distributed across 57 sites in grassland and oak savannah habitat in the Central Valley and surrounding foothills. The point count data set included 262 points in the primary focus area, of which 241 were in grassland and 21 were in oak savannah, and 249 points in the secondary focus area, of which 89 were in grassland and 160 were in oak savannah (Point Blue Conservation Science, unpublished data, see "Notes"). The point count method is an efficient method for estimating the valley oak woodland (VOW-S), coast oak woodland (COW-S), blue oak-foothill pine (BOP-S) and blue oak woodland (BOW-S) as oak savannah. We then calculated the area of grassland and oak savannah vegetation in the primary and secondary focus areas. Note that for grassland vegetation, pasture includes both irrigated and non-irrigated grasslands, as defined by the California Wildlife Habitat Relationships classification system (CDFW 2014). Although pastures may not have the same habitat value as other grassland types, we elected to include pasture in our assessment of grassland vegetation because it nonetheless provides habitat for nesting grassland birds, and is included in other statewide assessments of grassland vegetation (CPIF 2000). Notes: Conservation status designations include state bird species of special concern (SSC; Shuford and Gardali 2008), species ranked among the most vulnerable to climate change (CCV; Gardali et al. 2012), common birds in steep decline (CBSD; PIF 2012), U.S.-Canada species of conservation concern (UCC; PIF 2012), and near threatened (NT; BirdLife International 2014) 6 VOLUME 15, ISSUE 1, ARTICLE 6 relative abundance, richness, and diversity of a bird community (Ralph et al. 1993). Surveys began 15 minutes after local sunrise and concluded within 4 hours of sunrise. For each survey, the observer stood at each point and recorded every bird detected within a 5-minute period, noting the species, detection type (call, song, or visual), and distance from the observer. We identified each point count station as falling within either the primary or secondary focus areas, and, using vegetation data collected at each point count station, we identified -using the same criteria described above for classifying vegetation typeswhether each point count station was primarily in either grassland or oak savannah vegetation.
We then used the R packages "Distance" and "mrds" (Laake et al. 2015;Miller 2015) to conduct distance sampling analysis (Buckland et al. 2001). For this analysis, we fitted a standard set of recommended detection functions to the point count data for each focal species (Thomas et al. 2010) to estimate species densities (individuals ha -1 ) in each focus area and vegetation type, and to extrapolate the population sizes of each species across the estimated current extent of each vegetation type in each focus area. We also evaluated long-term population trends by examining trend estimates from Breeding Bird Survey (BBS) data within the Coastal California Bird Conservation Region (BCR 32) as reported by Sauer et al. (2014).
We could not use this approach to estimate current population sizes for the remaining four focal species: American Kestrel, Burrowing Owl, Northern Harrier, and Yellow-Billed Magpie, which had few or no detections in either the point count data or the BBS data. To estimate the population size for yellowbilled magpie, we used a published range-wide density estimate for the species (Crosbie et al. 2014), and extrapolated this density over the extent of oak savannah habitat in the primary focus area. We did not estimate Yellow-Billed Magpie population size in the secondary focus area because the species' range does not extend significantly into it (CWHR 1995

Central Valley Grassland and Oak Savannah Ecosystems
Because we used the focal species as indicators of the state of Central Valley grassland and oak savannah ecosystems, we evaluated the collective status (population size and trend) of all the focal species' populations. We applied a population status framework that is derived from general principles of conservation and population biology (Dybala et al. 2017, this volume). The framework is a hierarchy of four population size classes that mark milestones in the process of becoming a genetically robust, selfsustaining, and resilient wildlife population (Table 2). Population size thresholds for each population status were based on cross-taxa genetics studies, population viability analyses, and conservation biology theory for the orders of magnitude required for most vertebrate (primarily bird and mammal) populations. In addition, two modifiers indicate populations that are "steeply declining" or "resilient." Steeply declining populations are at high risk of extirpation or extinction regardless of population size, whereas resilient populations have multiple self-sustaining sub-populations in the planning area to guard against environmental catastrophes in one part of the range (see Dybala et al. 2017, this volume, for more information). We applied this framework to each focal species population in each focus area, and we

Long-Term Population Objectives
To achieve the long-term goal of genetically robust, self-sustaining, and resilient populations, our approach was to set population objectives for stable or increasing populations of each focal species in each focus area to be at least viable (>10,000 individuals) and preferably large (>50,000 individuals; Table 2). For the more common species that currently have viable or large populations, we defined longterm objectives to be reaching or maintaining large populations in each focus area. For the less common and special-status species that currently have small or very small populations, including Burrowing Owl, Loggerhead Shrike, and Yellow-Billed Magpie, we defined long-term objectives to be reaching or maintaining at least viable populations in each focus area. Because we lacked current density or population size estimates for American Kestrel and Northern Harrier, we assumed their populations to be small or very small, and we defined long-term population objectives to be reaching at least viable populations in each focus area.

Long-Term Density and Habitat Objectives
Because so much historical grassland and oak savannah vegetation has been lost from the Central Valley, we assumed many of the focal species populations are currently limited by available habitat. However, the condition of currently available grassland and oak savannah habitat has likely also been compromised as a result of habitat fragmentation, invasions of non-native species, and loss of vegetative cover in areas that have been heavily grazed by livestock (CPIF 2000;CalPIF 2002). Therefore, we assumed that the current densities of many of the focal species may be unusually low because of reduced habitat quality, and that habitat restoration and enhancement efforts would increase both the total area of habitat available to these species and their average densities. Consequently, our aim was to develop long-term habitat quality (density) and habitat quantity (hectares) objectives such that achieving both would result in meeting the long-term population objectives. Since some focal species occur in both grassland and oak savannah habitat, both of these habitat types contributed to their current population estimates. However, to set density objectives for habitat quality, we focused on the density in their preferred habitat association, i.e., the habitat type each species principally associates with (Table 1).
For the more common species that currently have large populations (>50,000 individuals), their longterm density objectives were simply to maintain

Population status Description
Proposed thresholds

Very small
Expected to be well below minimum viable population size (MVP), and at increased risk of inbreeding depression in the short term. < 1,000

Small
May be below MVP and vulnerable to extirpation through environmental and demographic stochasticity and long-term loss of genetic diversity. < 10,000

Viable
Expected to meet or exceed MVP, reducing vulnerability to environmental and demographic stochasticity and preserving genetic diversity. > 10,000

Large
Expected to be well above MVP, minimizing vulnerability to environmental and demographic stochasticity, preserving genetic diversity, and improving ability to maintain key ecological interactions and functions > 50,000

Steeply declining
Increased risk of extinction or extirpation until the causes of the decline are addressed, no matter the population size.

Resilient
Multiple viable or large populations to hedge against environmental catastrophes viable populations (>10,000) in more than one region 8 VOLUME 15, ISSUE 1, ARTICLE 6 their current average densities. For species that currently have population sizes (>10,000 individuals), we assumed increases in both density and habitat area would be required to be able to reach longterm density objectives of having large population sizes. However, information on historical breeding densities or the potential breeding densities that each species could reach is lacking. Thus, to balance increases in density with increases in habitat area, our initial hypothesis is that at least half the number of individuals required to reach a large population size could be achieved through increases in density in the preferred habitat, and the remainder through increases in preferred habitat area. Therefore, we calculated long-term density objectives (D obj , in individuals ha -1 ) for each species in each focus area as: where N p represents the current population size in the preferred habitat, N tot is the current total population size including both habitat types, and H p is the current area (in ha) of the preferred habitat. Then, assuming these density objectives are achieved, and that there is no change in the number of individuals in the secondary habitat (N s ), we calculated the total area of each species' preferred habitat that would still be required to achieve the long-term population objective of large (>50,000 individuals) in each focus area (H req ): From these species-specific estimates of the habitat area required (H req ), we determined the long-term habitat objective for each habitat type within each focus area as the largest of the individual H req estimates, such that the long-term habitat objectives were ultimately determined by the species that requires the greatest additional area of their preferred habitat. For the less common and specialstatus species that are currently small, very small, or unknown population sizes, for which we set long-term population objectives of reaching viable (>10,000 individuals), we defined long-term density objectives based on the densities required to achieve the population objective assuming the habitat objectives were achieved. Objectives for burrowing owl, northern harrier and yellow-billed magpie were not set in the secondary focus area since, historically, these species' breeding ranges only marginally included this region (CWHR 1995;Shuford and Gardali 2008).

Short-Term Density and Habitat Objectives
Because benchmarks can be useful for measuring progress toward achieving the long-term objectives, we present short-term (10-year) objectives as onetenth of the long-term habitat and density objectives. We assume that achieving these short-term objectives would be necessary to remain on track toward achieving the long-term objectives.

Current Status
Within the entire CVJV geography (primary and secondary focus areas combined), we estimated that there are currently 2.45 million ha (6.0 million acres) of grassland vegetation, where about 64% of the grasslands (1.57 million ha) are in the primary focus area and 36% (0.88 million ha) are in the secondary focus area (Figure 1; Table 3). Current focal species density estimates in grassland vegetation range from 0.003 individuals ha -1 for loggerhead shrike (in both the primary and secondary focus area) to 0.194 individuals ha -1 for western meadowlark in the primary focus area (Table 4).
For oak savannah vegetation, we estimated a current total of 0.7 million ha (1.8 million acres) distributed throughout both the primary and secondary focus areas combined (Figure 1; Table 3). The vast majority (94%) of the oak savannah habitat occurs within the secondary focus area, indicating that oak savannah habitat is extremely limited in the primary focus area (Figure 1; Table 3). In both focus areas combined, current focal species density estimates in oak savannah vegetation ranged from 0.003 individuals ha -1 for loggerhead shrike to 0.514 individuals ha -1 for western kingbird (Table 4).
Overall, densities were higher for focal species within their corresponding preferred breeding habitat. For example, western bluebird, a species principally   Note: In habitats where species densities are indicated with n/a, we assumed that densities were so low that they did not contribute significantly to regional population sizes. Densities could not be estimated from point count data for American kestrel or northern harrier because of few or no detections. b. For loggerhead shrike, we aggregated all detections across regions and habitat types (rather than stratifying across both) because of insufficient data.
c. For yellow-billed magpie, we used the density estimates reported in Crosbie et al. 2014. associated with oak savannah habitat, had density estimates in the primary focus area of 0.092 individuals ha -1 for oak savannah vegetation, but only 0.013 individuals ha -1 for grassland vegetation (Table 4).
Extrapolating the estimated vegetation totals across the current average density estimates of the focal species (all but American kestrel and northern harrier), the breeding population size estimates for each focus area ranged from 2,650 individuals (95% CI: 2,020-3,470; very small) for yellow-billed magpie to 309,330 individuals (95% CI: 235,760-406,040; large) for western meadowlark, both in the primary focus area (Table 5). We also estimated small or very small populations for burrowing owl and loggerhead shrike in both focus areas ( Figure 2). Further, according to BBS trend estimates for BCR 32, two-thirds of the focal species are significantly declining, and two of the focal species, horned lark and burrowing owl, meet the criteria for steeply declining (Table 2; Appendix A). Fitting these population size and trend estimates into the population status framework, we found that only two of the five species associated with grassland habitat (40%) and four of the seven species associated with oak savannah habitat (57%) are currently Note: In habitats where species densities are indicated with n/a, we assumed that densities were so low that they did not contribute significantly to regional population sizes. Densities could not be estimated from point count data for American kestrel or northern harrier as a result of few or no detections. populations -and calculated that an additional 34,329 ha of oak savannah habitat would be required for all these species to reach their long-term population objectives of large. The long-term total habitat objective for oak savannah in the primary focus area is 79,942 ha, a 75% increase over the current extent of oak savannah habitat. In contrast, four species already had large populations in the secondary focus area, and the three remaining species all had small, unknown, or not applicable population sizes ( Figure 2). Thus, we set a long-term habitat objective to maintain the existing 676,666 ha of oak resilient, with viable (>10,000 individuals) or large (>50,000 individuals) populations in each focus area ( Figure 2).

Long-Term Density and Habitat Objectives
Of the five focal species whose preferred habitat is grassland, we set long-term density objectives of maintaining their current densities (Table 6A) for horned lark and western meadowlark, which already have large populations in both focus areas (Figure 2). Of the three remaining species, burrowing owl and northern harrier had small, very small, or unknown population sizes, such that grassland habitat objectives were ultimately determined by the density objectives for the fifth species: grasshopper sparrow. From their current average densities of 0.017 (0.009-0.034) and 0.016 (0.006-0.044) individuals ha -1 in the primary and secondary focus regions, respectively (Table 4A), we originally calculated grasshopper sparrow density objectives of 0.025 and 0.039 individuals ha -1 , which is still within the 95% confidence interval range of our current density estimates. However, even these longterm objectives were much lower than observed density estimates for grasshopper sparrows in other regions, which range to well over 0.5 individuals ha -1 (Collier 1994;Vickery 1996;CPIF submitted, see "Notes"). Relying on such low density objectives would have resulted in very large grassland habitat objectives. Therefore, we raised the long-term density objective to 0.05 individuals ha -1 , just above the upper confidence interval of our calculated density of 0.044 individuals ha -1 (Table 4). Consequently, we estimated that no additional grassland area was necessary to achieve the long-term population objectives in the primary focus area; the long-term grassland habitat objective is to maintain the current 1,567,257 ha and ensure that no net loss occurs (Table 3). In the secondary focus area, the long-term grassland habitat objective is to increase the extent by 41,831 ha to a total of 921,821 ha, a 5% increase.
Of the seven focal species whose preferred habitat is oak savannah, only western kingbird currently has a large population in the primary focus area. Thus, we set higher long-term density objectives for acorn woodpecker, lark sparrow, and western bluebird -the three species that currently have viable Projected population status if long-term (100-year) population objectives are achieved. A status of n/a in one of the focus areas means the species is not expected to breed in that focus area. A status of "unknown" means the current population size or trend is currently unknown and the species is assumed not to be viable, large, or resilient. Thus, the calculation of % viable, large, or resilient represents a minimum value.
12 VOLUME 15, ISSUE 1, ARTICLE 6 savannah in the secondary focus area and to ensure that no net loss occurs (Table 3). For all remaining species with currently small, very small, or unknown population sizes, we set long-term density objectives that would ensure they could reach their long-term population objectives of viable if the long-term habitat objectives were achieved. However, we note that for species with unknown population sizes, it is not known whether these density objectives are higher or lower than current average breeding densities.

Short-Term Density and Habitat Objectives
As a milestone for achieving the long-term objectives within a 100-year time-frame, we set short-term (10-year) objectives for each region equal to onetenth of the improvement required to reach the longterm population, density, and habitat objectives. The short-term habitat objective for grassland habitat is to maintain the same extent in area for the primary focus area, and increase grassland habitat by 4,183 ha in the secondary focus area (Table 3B). For oak savannah habitat, the short-term objective in the primary focus area is to increase the extent by 3,432 ha, and to maintain the current extent of 676,666 ha in the secondary focus area (Table 3B).

DISCUSSION
This effort represents the first attempt at developing population and habitat objectives for valley grassland and oak savannah landbirds in the Central Valley and its surrounding foothills. We modeled our approach using focal species as indicators of grassland and oak savannah ecosystem condition on the conservation planning efforts for riparian landbirds in the Central Valley (Dybala et al. 2017, this volume). Here, we selected 12 focal species that represent a range of habitat attributes found in grassland and oak savannah ecosystems, and we developed population, density, and habitat objectives to meet the longterm goal of genetically robust, self-sustaining, and resilient focal species populations. Protecting, restoring, and managing Central Valley grassland and oak savannah ecosystems to support this set of focal species would potentially benefit other wildlife that also depend on these ecosystems, including species that are declining or at-risk such as Swainson's Hawk (Buteo swainsoni) and the San Joaquin kit fox (Vulpes macrotis mutica; Stromberg et al. 2007).
These conservation activities will also contribute to the prosperity and quality of life in the Central Valley by providing for carbon sequestration and pollinator refugia (Havstad et al. 2007;Kroeger et al. 2009;Chaplin-Kramer et al. 2011). In the context of rangeland conservation needs, conserving grasslands and oak savannah ecosystems would also contribute to food and fiber production (Havstad et al. 2007;Cameron et al. 2014). To achieve these objectives, it will be necessary to protect and enhance existing habitat and restore habitat where land is currently in other uses. Depending on the site and focal species, habitat enhancement efforts might include the removal of noxious weeds, such as yellow star-thistle (Centaurea solstitialis), and encouraging regeneration of blue oaks (Quercus douglasii) and greater cover of native bunchgrasses, such as purple needlegrass (Stipa pulchra). Best strategies for achieving these enhancement objectives would be site-and projectspecific, but they could include prescribed burning, prescribed grazing, and native plantings. Restoration, on the other hand, would establish (or re-establish) grassland and oak savannah vegetation through active seeding and plantings, and would likely require long-term vegetation management until the plants were well established.
There are examples of small-scale grassland restoration projects in the Central Valley (<100 ha), though they appear to have had limited success in supporting breeding grassland birds, perhaps because these restored grasslands were smaller than the patch size requirements for the grassland birds being monitored (DiGaudio et al. 2009;Young and DiGaudio 2011). Future restoration projects should be strategically located to improve habitat connectivity and build upon existing grassland and oak savannah patches. Given that each of the focal species has its own distinct set of habitat requirements (e.g., horned larks and burrowing owls prefer short-stature grasslands, whereas meadowlarks prefer taller grasslands), restoration and enhancement projects should consider what their target management species are relative to their habitat requirements, and attempt to create habitat mosaics across the landscape to accommodate the needs of multiple species.
Managed livestock grazing could play a significant role in enhancing grassland and oak savannah habitat for birds, especially given that the vast majority of California's grasslands and oak savannahs are currently used for livestock production (Stromberg et al. 2007). There is still much to learn, however, about optimizing rangeland management and livestock grazing practices to benefit birds and ecosystem services. Recommendations have been put forward to improve habitat conditions for the grassland and oak savannah focal species; however, most recommendations are hypothetical, and evaluating their effectiveness will require further testing and validation. For example, grasshopper sparrows are associated with perennial bunchgrasses; therefore, increasing perennial grass cover should increase grasshopper sparrow density (Vickery 1996). Such specific recommendations can be found for each focal species in the California Partners in Flight Grassland Bird Conservation Plan (CPIF, unreferenced, see "Notes").
Achieving the long-term habitat objectives of increasing oak savannahs in the primary focus area by 75% and grasslands in the secondary focus area by 5% will be challenging, especially since the expansion of one habitat type should not come at the expense of replacing or converting another habitat type. For example, oak savannah should not replace existing grasslands, nor should oak forests be cleared to create grasslands. Given this planning consideration, more research is necessary to determine whether achieving these habitat objectives would be feasible through the conversion of existing agricultural lands (i.e., row crops, orchards, and vineyards) or other lands that have limited ecological value in their current condition.

Measuring Success
We expect that restoring grassland and oak savannah vegetation, particularly when strategically located to reduce fragmentation, will result in an improvement in species densities, and, in turn, help achieve the population objectives. Measuring success will require grassland and oak savannah breeding bird communities and their habitat to continue to be monitored, to detect changes in their density and overall population sizes. A standardized bird survey effort at randomized and spatially balanced locations throughout the primary and secondary focus areas would provide a robust estimate of average species densities and how they change through time. Paired with efforts to track grassland and oak savannah restoration and enhancement projects in each focus area, changes in total population size and status -and thus progress toward the long-term population objectives -can be monitored.
In our approach to setting objectives such that each focus area will support genetically robust, selfsustaining, and resilient focal species populations, we focused our efforts only on the proportion of the population that occurs in grassland and /or oak savannah habitat. These habitats likely cover the vast majority of the population for some of the grassland-dependent species, namely burrowing owl, grasshopper sparrow, horned lark, and western meadowlark. However, for other focal species, some proportion of their populations also occurs in other habitat types, which we did not include in our efforts to estimate their current population sizes: these include oak woodland, riparian, and, for some species, agricultural settings (e.g., Northern Harriers use alfalfa fields for foraging and nesting; Shuford and Gardali 2008). For these focal species, these other habitats contribute to their regional populations, which means we may have over-estimated the density and habitat objectives for grasslands and oak savannahs necessary to achieve the population objectives. Having a more detailed understanding of each focal species' regional population, and how the populations are distributed across different habitat types within each region, would allow us to formulate more precise population objectives specific to grassland and oak savannah habitat.
In addition, we assumed that many focal species' densities are currently reduced as a result of extensive fragmentation and degradation of grassland and oak savannah habitat in the Central Valley, and our density objectives therefore represent hypotheses for densities that can -with habitat restoration and enhancement -be achieved in the Central Valley. Since we did not assess habitat condition from each point count location, we were unable to evaluate how habitat condition affected focal species density. However, the large confidence intervals around many of the current density estimates may be explained, in part, by varied habitat conditions throughout the study area. These hypotheses need to be tested by changes in average species densities being monitored as grassland and savannah restoration and enhancement efforts continue. If these density objectives prove to be too low, achieving the longterm population objectives may not require as much additional habitat, and the habitat objectives will need to be revised downward. On the other hand, if these density objectives prove to be unreasonably high, the habitat objectives may need to be revised upward.
The long-term population objectives were set for each focus area rather than for the Central Valley as a whole, in part because the long-term goal was to achieve resilient populations with viable or large subpopulations in each focus area, and, in part, because we hypothesize that each area can still support large populations. Since the primary focus area generally encompasses the valley floor, and the secondary focus area encompasses the surrounding foothills, it seems reasonable to assume that there would be inherent ecological differences between the two areas, including their respective capacities for supporting focal species populations. Continued monitoring of grasslands and oak savannahs in each area, and modeling changes going forward, will be important in determining whether the potential capacity of any area has changed, and whether the population objectives should be revised.
The long-term population objectives are based on general hypotheses for the orders of magnitude required to achieve genetically robust, self-sustaining, and resilient populations (Dybala et al. 2017, this volume). They are based on principles of conservation biology and are derived from the best available data, but may require revision if new species-or population-specific information becomes available. For example, new population viability analyses could indicate that smaller population sizes can be considered robust, or new information about important ecological functions could indicate the minimum population size that should be considered large. Finally, achieving these population objectives is never a guarantee of population persistence. It will be important to continue to monitor bird populations, test our assumptions, reduce uncertainties, and revise our estimates of what is required to achieve genetically robust, self-sustaining, and resilient populations.

CONCLUSIONS
The Central Valley has lost about 60% of its historical valley grassland habitat, and an unknown amount of its oak savannah habitat. Much of the https://doi.org/10.15447/sfews.2017v15iss1art6 remaining habitat is degraded or at risk of conversion to urban or intensive agricultural development. The remaining area currently supports mostly large and viable populations, though some species have steeply declining populations or have small to very small populations that may be vulnerable to extirpation in the long-term. To achieve our longterm goal of genetically robust, self-sustaining, and resilient grassland and oak savannah breeding bird populations in the Central Valley, we estimate that the extent of grassland habitat needs to increase by 2% (41,831 ha), and savannah by 5% (34,329 ha). Additionally, there is a need to increase species densities in most existing habitat by enhancing current habitat conditions. By working toward achieving these population, density, and habitat objectives through threat abatement, habitat restoration, and habitat enhancement, we expect grassland and oak savannah ecosystem function and resilience to improve, which will, in turn, benefit other wildlife populations and the people of the Central Valley and beyond.