Monitoring of the MacNeill’s Sootywing and its Habitats 2020 Annual Report

April 2021

Work conducted under LCR MSCP Work Tasks D14 and F6 Lower Colorado River Multi- Conservation Program Steering Committee Members

Federal Participant Group California Participant Group

Bureau of Reclamation California Department of Fish and Wildlife U.S. Fish and Wildlife Service City of Needles National Park Service Coachella Valley Water District Bureau of Land Management Colorado River Board of California Bureau of Indian Affairs Bard Water District Western Area Power Administration Imperial Irrigation District Los Angeles Department of Water and Power Palo Verde Irrigation District Arizona Participant Group San Diego County Water Authority Southern California Edison Company Arizona Department of Water Resources Southern California Public Power Authority Arizona Electric Power Cooperative, Inc. The Metropolitan Water District of Southern Arizona Game and Fish Department California Arizona Power Authority Central Arizona Water Conservation District Cibola Valley Irrigation and Drainage District Nevada Participant Group City of Bullhead City City of Lake Havasu City Colorado River Commission of Nevada City of Mesa Nevada Department of Wildlife City of Somerton Southern Nevada Water Authority City of Yuma Colorado River Commission Power Users Electrical District No. 3, Pinal County, Arizona Basic Water Company Golden Shores Water Conservation District Mohave County Water Authority Mohave Valley Irrigation and Drainage District Native American Participant Group Mohave Water Conservation District North Gila Valley Irrigation and Drainage District Hualapai Tribe Town of Fredonia Colorado River Indian Tribes Town of Thatcher Chemehuevi Indian Tribe Town of Wickenburg Salt River Project Agricultural Improvement and Power District Unit “B” Irrigation and Drainage District Conservation Participant Group Wellton-Mohawk Irrigation and Drainage District Yuma County Water Users’ Association Ducks Unlimited Yuma Irrigation District Lower Colorado River RC&D Area, Inc. Yuma Mesa Irrigation and Drainage District The Nature Conservancy

Other Interested Parties Participant Group

QuadState Local Governments Authority Desert Wildlife Unlimited

Lower Colorado River Multi-Species Conservation Program

Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Prepared by: Jeff Hill, Wildlife Group Nathan Rudd, Adaptive Management Group

Lower Colorado River Multi-Species Conservation Program Bureau of Reclamation Lower Colorado Basin Boulder City, Nevada http://www.lcrmscp.gov April 2021

Hill, J. and N. Rudd. 2021. Monitoring of the MacNeill’s Sootywing Skipper and its Habitats, 2020 Annual Report. Lower Colorado River Multi-Species Conservation Program, Boulder City, Nevada.

ACRONYMS AND ABBREVIATIONS

AIC Akaike’s information criterion AICc Akaike’s information criterion adjusted for small sample size

CI confidence interval Cibola NWR Cibola National Wildlife Refuge CVCA Cibola Valley Conservation Area ha hectare(s) km kilometer(s)

LCR MSCP Lower Colorado River Multi-Species Conservation Program n sample size

RH relative humidity

SE standard error sootywing MacNeill’s sootywing skipper (Pholisora gracielae = gracielae [MacNeill])

UTM NAD 83 Universal Transverse Mercator North American datum of 1983

Symbols

≈ approximately

> greater than

x multiplied by

% percent ± plus or minus

rs Spearman’s Rho

CONTENTS

Page

Introduction ...... 1 Methods...... 1 Data Collection Method ...... 1 Survey ...... 1 Measurements ...... 2 Abiotic ...... 2 Quailbush ...... 2 Floral Metrics ...... 2 Sootywing ...... 3 Analysis...... 3 Sample Sizes and Preliminary Comparisons ...... 3 Correlations ...... 4 Logistic Regression ...... 4 Software Used ...... 5 Study Areas ...... 5 Big Bend Conservation Area ...... 6 Beal Lake Conservation Area ...... 8 Palo Verde Ecological Reserve ...... 10 Cibola Valley Conservation Area ...... 12 Cibola National Wildlife Refuge Unit #1 Conservation Area ...... 14 Pretty Water Conservation Area ...... 16 Cibola National Wildlife Refuge-Island Unit ...... 18 Hart Mine Marsh ...... 20 Hunters Hole ...... 22 Results ...... 24 Sootywing Detections, 2015–2019 ...... 24 Predictors of Sootywing Presence ...... 26 Discussion ...... 28 Literature Cited ...... 31

Tables

Table Page

1 Sites sampled for adult and immature stages of sootywings ...... 5 2 Spearman’s rank correlation coefficients (ρ) of quailbush attributes and counts of sootywing eggs and adults ...... 27 3 Odds ratios for predictors from the best logistic regression model on egg presence ...... 28

i Figures

Figure Page

1 Big Bend Conservation Area managed acreage through fiscal year 2019...... 7 2 Beal Lake Conservation Area managed acreage through fiscal year 2019...... 9 3 Palo Verde Ecological Reserve managed acreage through fiscal year 2018...... 11 4 Cibola Valley Conservation Area managed acreage through fiscal year 2019...... 13 5 Cibola Wildlife Refuge Unit #1 managed acreage through fiscal year 2020...... 15 6 Pretty Water Conservation Area managed acreage through fiscal year 2019...... 17 7 Cibola NWR-Island Unit, fiscal year 2017...... 19 8 Hart Mine Marsh managed acreage through fiscal year 2019...... 21 9 Hunters Hole managed acreage through fiscal year 2019...... 23 10 Total surveys conducted per site, 2015–19...... 24 11 Mean adult sootywings detected per survey, 2015–19...... 25 12 Mean eggs detected per quailbush surveyed, 2015–19...... 25 13 Scatter plot matrix of ranked quailbush attributes and counts of eggs and adults with density histograms and linear fits for data collected at conservation areas, 2015–19...... 26 14 Predicted probability (with 95% confidence band or interval) of egg presence versus leaf density (left) and quailbush height (feet) (right)...... 28

Attachments

Attachment

1 Survey Area Maps

ii

INTRODUCTION

The MacNeill’s sootywing skipper (Pholisora gracielae = Hesperopsis gracielae [MacNeill]) (hereafter sootywing) is a small (wingspan 0.79 to 1.25 inches, 20 to 32 millimeters), dark-colored skipper endemic to the lower Colorado River system (Pratt and Wiesenborn 2011). Larvae of the sootywing can only complete development on quailbush (Atriplex lentiformis) (Wiesenborn and Pratt 2008). A variety of other plant species are used for nectar by adult sootywings (Wiesenborn and Pratt 2010). Alkali heliotrope (Heliotropium curassavicum) and western purslane (Sesuvium verrucosum) are considered important nectar sources for sootywing habitat creation (Wiesenborn and Pratt 2010).

This skipper is the only invertebrate covered by the Lower Colorado River Multi- Species Conservation Program (LCR MSCP). The LCR MSCP is expected to facilitate a balance between anthropogenic use of Colorado River resources and the conservation of native species and their habitats. Information contained within this report concerns sootywing presence in honey mesquite (Prosopis glandulosa) plots as well as Fremont cottonwood (Populus fremontii) and Goodding’s willow (Salix gooddingii) (hereafter cottonwood-willow) plots containing quailbush in LCR MSCP conservation areas. Quailbush attributes at various locations were examined as well as associations with the egg and larval life stages of sootywings.

METHODS Data Collection Method

For all seasons, data were recorded on a mobile electronic field form developed by the LCR MSCP (LCR MSCP 2015). From 2015 to 2017, data and approximate locations were taken using a Trimble Juno Global Positioning System unit running the field forms using the software Terrasync. Data were downloaded from the Trimble units and processed in Microsoft Access. In the 2018–19 survey seasons, data were collected on an iPhone or iPad running ArcGIS Survey123 software and downloaded and processed to ArcGIS Online.

Survey

Two people spent a total of 5 minutes searching each selected quailbush for sootywing eggs and caterpillars. After quailbush characteristics and immature sootywing observations were completed, timing for the next stop was initiated. Any adult sootywings that were encountered were also enumerated. Identification of various life stages of sootywings utilized information in Nelson et al. (2015).

1 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

If eggs or caterpillars were found at additional quailbush during a survey, information was also collected at these supplementary plants. Quailbush damage associated with presumptive larval sootywings and invertebrates other than sootywings were also noted.

The timing of monitoring was designed to avoid the hottest times of the day (approximately 1:00 to 3:00 p.m.). Previous monitoring (e.g., Nelson et al. 2015; Pratt and Wiesenborn 2009) suggests that adult sootywings are difficult to find during the warmest time of day perhaps because they are avoiding activity and seeking shade within quailbush.

Measurements Abiotic Soil moisture at the base of each plant was measured. From 2015 to 2017, moisture (percent saturation relative to field capacity) was measured with a Kelway soil moisture tester. From 2018 to 2019, moisture was measured with a Model HB-2 Campbell Scientific HydroSense II soil-water sensor. Windspeed (kilometers per hour), air temperature (degrees Fahrenheit), relative humidity (RH), and lux were collected at the start and end of each sampling occasion. Windspeeds were measured with a Kestral 3000 Wind Meter Anemometer Meter (±3% accuracy) and recorded on the Beaufort scale (an empirical measure of wind strength ranging from calm [force 0] to hurricane [force 12]). Air temperature and RH were measured with a hand-held Extech Easy View 20 Hydro-thermometer (RH range = 10 to 95% with 0.1% resolution and basic accuracy of ±3% [30 to 95% RH] and ±5% [105 to 30% RH]). Lux was measured with a hand-held Extech 401025 light probemeter (resolution of 1 lux with 5% accuracy). Temperature and RH were also recorded each time an adult sootywing was detected.

Quailbush Quailbush were measured for height, width (nearest 0.1 foot), an estimated percentage of dry or absent leaves (an indicator of plant lushness visually agreed upon by two observers), leaf density (low, medium, high), and leaf growth (small, medium, large).

Floral Metrics Estimates of sootywing habitat quality also included floral (nectar) measurements. Floral abundance by plant species in the immediate environment was qualitatively noted, where 0 = none, 1 = scarce (flowers rarely encountered), 3 = common (flowers often observed), and 5 = abundant (floral abundance unlikely to be

2 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

limiting). The floral index consisted of the sum of the value recorded for each plant species: If three separate plant species all had abundant flowers, the plot score would be 15. This index favors both floral abundance and diversity of flower sources. The presence of honey mesquite and screwbean mesquite (Prosopis pubescens) was documented at the end of each survey.

Sootywing

During sootywing monitoring in conservation areas populated by quailbush, information was collected on five quailbush randomly selected by time (timer set at 5 minutes) while walking through the survey site. When quailbush were sparse, information was collected at each quailbush encountered until the five quailbush limit was reached. The set stopping time allowed for selection of a quailbush dependent only upon the previous path taken through the habitat. The walk was not truly random, however, because the route through the environment was directed by the density of quailbush and the presence of nectar plants. Surveys began in April and continued each month until sootywings were detected or through August. Once sootywings were detected in a survey, the conservation area was deemed occupied, and no additional surveys were required at that conservation area that year.

Behaviors of detected adult skippers were recorded as flying, perching, basking (wings open), nectaring (probing of flower with proboscis), puddling, mating, and ovipositing. A sootywing’s sex was also recorded when distinguished; females are identified by paler and more mottled forewings and typically by a larger body size compared to males.

Analysis Sample Sizes and Preliminary Comparisons As measurements taken changed through the years, the sample sizes analyzed varied. Sootywing presence was compared to the following measurements: percent dry quailbush, quailbush area (height x width), caterpillar damage, and soil moisture (n = 267); leaf density and leaf growth (n = 181); and weather measurements were taken once per survey (ambient temperature, relative humidity, and cloud cover ) (n = 54). The average of hourly soil moisture measurements taken over the previous 2 weeks from nearby long-term soil moisture sensors (n = 4) was also compared with mean quailbush attributes and sootywing counts to determine if data from higher quality equipment over a longer time period was more highly correlated with sootywing abundance. Preliminary analyses indicated no reliable relationship between average soil moisture from long-term sensors and sootywing abundance or other variables at such a small sample size; therefore those results are not included in this report.

3 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Correlations Scatter plots and Spearman’s rank correlations were used to identify and measure associations between quailbush attributes and counts of sootywing eggs and adults. Variables were rank transformed for scatter plots to be consistent with the correlation analysis. The Spearman’s correlation was chosen over the Pearson’s correlation because most habitat descriptors were ordinal. The combined influence of quailbush attributes on egg presence was examined with multiple logistic regression. Eggs were considered a more sensitive indicator of sootywing preference for quailbush than adults, as adults are attracted to areas with abundant nectar sources. Also, the method of counting adults varied over years; they were not counted at measured quailbush until 2017. Since egg counts did not vary sufficiently over time to be informative (range 0–9; 70% of quailbush had no eggs), egg presence was modeled.

Logistic Regression Backward and forward stepwise logistic regression was used to select the best model based on changes in Akaike’s information criterion (AIC) (Burnham and Anderson 2002) as model predictors were removed or added. Models were fit only to observations with no missing values for any predictors (n = 181). Quailbush height (feet) and area (square feet) were highly correlated, so separate stepwise selections were done for models including each measure of plant size. The “full” model for backward selection included plant size, caterpillar damage, soil moisture, leaf density, and leaf size, along with all two-way interactions. Other attributes less correlated with eggs were not considered. Models with higher level interactions (> 2 predictors) did not converge and therefore could not be examined. The base model was a null model with no predictors. The final model from each stepwise run with the lowest AICc score (AIC corrected for small sample size) was considered the best overall model for the data considered in this report.

An important assumption of logistic regression is independence of error terms, which is often unrealistic for a nested sampling design (i.e., quailbush attributes and sootywing abundance/presence were likely more similar within, than among sites, as well as within years than over years). Violating this assumption can result in underestimates of standard errors (SEs), inflated p-values, and biased confidence intervals (CIs). Durbin-Watson tests supported autocorrelated errors, so robust SEs were calculated using the type HC3 sandwich estimator (ref). Odds ratios with 95% CIs based on robust SEs are reported for the final model as well as graphs of predicted probability for each predictor (with 95% CIs).

4 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Software Used Statistical analyses were done in R (v. 4.0.2, R Core Team 2020). Functions scatterplotMatrix and correlation (‘correlation’ package, Makowski et al. 2019) were used for scatter plots and correlation analysis, which included t-tests of significance for correlation parameters uncorrected for multiple tests. Function stepAIC from the MASS package (Venables and Ripley 2002) was used for logistic model selection; summ and plot_model (‘sjPlot’ package, Lüdecke 2020) were used for calculating robust SE and graphing predicted probability of egg presence for varying levels/values of model predictors. performance_hosmer from the performance package (Lüdecke et. al 2020) provided the Hosmer- Lemeshow goodness of fit test. AICc (‘AICcmodavg’ package, Mazerolle 2020) was used to calculate the small-sample value of AIC.

STUDY AREAS

Monitoring of both adult and immature stages of sootywings occurred at plots containing quailbush in LCR MSCP conservation areas. Sites, locations and years surveyed are presented in table 1. Survey area maps are included in attachment 1.

Table 1.—Sites sampled for adult and immature stages of sootywings (Sites are arranged upstream to downstream along the Colorado River.) Coordinates (UTM NAD 83) Site Northing Easting Years surveyed Big Bend Conservation Area 3887331 714174 2015, 2016, 2017 Beal Lake Conservation Area 3851153 725478 2015 Palo Verde Ecological Reserve-Phase 1 3730595 728268 2015, 2016, 2018, 2019 Palo Verde Ecological Reserve-Phase 4 3730972 730630 2015, 2016, 2019 Palo Verde Ecological Reserve-Phase 5 3731560 731043 2019 Palo Verde Ecological Reserve-Phase 6 3732430 731009 2015, 2016 Palo Verde Ecological Reserve-Phase 8 3734140 731057 2019 Cibola Valley Conservation Area-Phase 2 3698616 717014 2018 Cibola Valley Conservation Area-Phase 3 3698072 714138 2019 Cibola Valley Conservation Area-Phase 4 3698908 714203 2015, 2016 Cibola Valley Conservation Area-Phase 5 3698953 715007 2018, 2019 Cibola Valley Conservation Area-Phase 6 3699314 715003 2018 Cibola Valley Conservation Area-Phase 9 3695342 713895 2019 Cibola National Wildlife Refuge Unit #1 Conservation Area-Crane Roost 3694126 715384 2016 Pretty Water Conservation Area 3691888 713790 2017, 2018, 2019 Cibola National Wildlife Refuge -Island Unit 3687202 714368 2015, 2016, 2017 Hart Mine Marsh 3685867 717854 2015, 2016 Hunters Hole South Cell 3600220 706562 2015

5 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Big Bend Conservation Area

The Big Bend Conservation Area is located in Nevada 5 miles (8 kilometers [km]) south of Laughlin, Nevada, along the Needles Highway (figure 1). The site is within Reach 3. The site includes 15 acres of backwater as well as 15 acres of habitat comprised of a marshy strip of cattails (Typha spp.) leading into a drier strip of arrowweed (Pluchea sericea) and baccharis (Baccharis salicifolia) (figure 1). The upland mesquite portion of the property has 15 acres planted with honey and screwbean mesquite. There are a few (≈3) small quailbush found in the upland portion of the conservation area.

6 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 1.—Big Bend Conservation Area managed acreage through fiscal year 2019.

7 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Beal Lake Conservation Area

The Beal Lake Conservation Area is 100 acres adjacent to Beal Lake and Topock Marsh, inside the Havasu National Wildlife Refuge on the Arizona side of the Colorado River (figure 2). The site is within Reach 3. The site contains 119 acres of cottonwood-willow habitat and 225 acres backwater habitat. It is a two-phase habitat creation project that was initiated in spring 2003. The site was planted with cottonwood-willow, coyote willow (Salix exigua), honey mesquite, and screwbean mesquite. Arrowweed and some baccharis have begun to fill in the open areas and edges of most of the plots in the site. Quailbush grow on the edges of the planted cottonwood-willow adjacent to the willow marsh as well as north of the conservation area boundary.

8 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 2.—Beal Lake Conservation Area managed acreage through fiscal year 2019.

9 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Palo Verde Ecological Reserve

The Palo Verde Ecological Reserve is a conservation area located about 5 miles (8 km) north of Blythe, California, along the California side of the Colorado River (figure 3). The site is within Reach 4 and encompasses 1,300 acres (526 hectares [ha]). The conservation area was planted in eight different phases, with one phase planted every year from 2006 through 2014. In spring 2006, a 61-acre (25-ha) nursery (Phase 1) was planted. In spring 2007, Phase 2 was planted with 80 acres (32 ha) of cottonwood-willow, and other riparian plants. Phase 3 was planted in spring 2008 and is also planted with cottonwood-willow habitat types. Phase 4 was planted in 2009 and contains mostly cottonwood-willow, with one plot consisting of a mix of honey mesquite, quailbush, and a mix of native grasses. Phases 5, 6, and 7 were planted in 2010, 2011, and 2012 with cottonwood-willow habitat. Phase 8 was planted in 2013 with 38 acres (15 ha) of honey mesquite and alkali sacaton (Sporobolus airoides), and it contains cottonwoods that seeded naturally.

The area surveyed in Phase 1 consisted of honey mesquite and quailbush of varying densities and encompassed both lush and completely desiccated plants. The areas surveyed in Phases 4, 5, and 6 occurred along strips of quailbush growing beside planted cottonwood-willow habitat. Surveys in Phase 8 occurred in planted mesquite habitat that had been colonized by cottonwoods and quailbush.

10 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 3.—Palo Verde Ecological Reserve managed acreage through fiscal year 2018.

11 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Cibola Valley Conservation Area

The Cibola Valley Conservation Area is located in Arizona adjacent to the Colorado River, about 15 miles (24 km) south of Blythe, California (figure 4). The site is within Reach 4 and encompasses 1,235 acres (500 ha). Phase 1 was planted in spring 2006 and contains a 22-acre (9 ha) nursery and a 64-acre (26 ha) area of cottonwood-willow habitat. Phase 2 comprises 71 acres (29 ha) and was planted in spring 2008. It consists mostly of cottonwood-willow habitat, with one small area of honey mesquite and quailbush. Phase 3 was planted in spring 2007 and contains over 80 acres (32 ha) of cottonwood-willow planted in different combinations. Phase 3 also includes 11 acres (4 ha) of baccharis mixed with some cottonwood-willows. Phase 4 was planted in 2009 with 245 acres (99 ha) of honey mesquite and quailbush. Phase 5 was planted in 2010 with 71 acres (29 ha) of honey mesquite and quailbush. Phase 6 was planted in 2011 with 89 acres (36 ha) of honey mesquite. Phase 9 was planted in 2017 with 76 acres (28 ha) of a mosaic of cottonwood-willow and honey mesquite. Phases 3, 4, 5, 6, and 9 were surveyed. The areas surveyed in the Cibola Valley Conservation Area consisted of quailbush and honey mesquite planted in rows.

12 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 4.—Cibola Valley Conservation Area managed acreage through fiscal year 2019.

13 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Cibola National Wildlife Refuge Unit #1 Conservation Area

The Cibola National Wildlife Refuge (Cibola NWR) is located approximately 30 miles (48 km) south of Blythe, California, along 12 miles (19 km) of the LCR in Arizona and California (figure 5). The site is within Reach 4. The Cibola NWR is divided into six management units, of which the Cibola National Wildlife Refuge Unit #1 Conservation Area comprises approximately 949 acres. Area #1 was planted from 1999 to 2009 and consists 155 acres of cottonwood-willow habitat. Large quailbush are sporadically growing in along some edges in the site. Crane Roost was planted from 2008 to 2009 and consists of 154 acres of cottonwood-willow habitat. There is a strip of large, lush quailbush growing along the north edge and the edges of the roads intersecting the site. Area # 2 was planted with 339 acres of cottonwood-willow habitat from 2013 to 2016. There are sparse quailbush growing in along the open areas adjacent to irrigation canals in the northern part of the site.

14 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 5.—Cibola Wildlife Refuge Unit #1 managed acreage through fiscal year 2020.

15 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Pretty Water Conservation Area

The Pretty Water Conservation Area consists of approximately 566 acres (225 ha) on the Cibola NWR, located in California between River Miles 95 and 97 (figure 6). The site was planted with honey mesquite habitat in 2015. The area surveyed was a mix of arrowweed and mature saltcedar (Tamarix spp.), with sporadic quailbush present along the road and within the site.

16 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 6.—Pretty Water Conservation Area managed acreage through fiscal year 2019.

17 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Cibola National Wildlife Refuge-Island Unit

The Cibola NWR-Island Unit is located approximately 22 miles (35 kms) south of Blythe, California, at the south end of the refuge (figure 7). The site is within Reach 5. It consists of a number of areas planted with Fremont cottonwoods and honey mesquite. A fire at the site in 2011 resulted in a majority of the honey mesquite habitat being cleared. An area with natural recruitment of honey mesquite and quailbush at the south end of the site was sampled to assess post-fire colonization by sootywings.

18 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 7.—Cibola NWR-Island Unit, fiscal year 2017.

19 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Hart Mine Marsh

Hart Mine Marsh is located in the Cibola NWR, approximately 22 miles (35 km) south of Blythe, California (figure 8). It contains 163 acres of marsh restored between 2009 and 2011. The site is located at the southern limit of Reach 4 (figure 5). The vegetation consists of southern cattail (Typha domingensis), chairmaker’s bulrush (Schoenoplectus americanus), California bulrush (Schoenoplectus californicus), common spikerush (Eleocharis palustris), softstem bulrush (Schoenoplectus tabernaemontani), and inland saltgrass (Distichlis spicata) with quailbush and honey mesquite. There is a patch of quailbush north of the conservation area boundary.

20 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 8.—Hart Mine Marsh managed acreage through fiscal year 2019.

21 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Hunters Hole

Hunters Hole is a conservation area located near the community of San Luis, 18 miles (29 km) south of Yuma, Arizona (figure 9). The site is within Reach 7. It was added to the LCR MSCP in 2011. The site was planted in spring 2012 with approximately 44 acres of cottonwood-willow habitat consisting of cottonwood- willow, honey mesquite, and marsh. A number of fires have occurred at Hunters Hole throughout the years, which periodically open up the habitat for a dense layer of grasses to grow in before being shaded out by canopy regrowth.

22 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 9.—Hunters Hole managed acreage through fiscal year 2019.

23 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

RESULTS Sootywing Detections, 2015–2019

A total of 54 surveys were conducted from 2015 to 2019 (figure 10). The highest mean adult detections per survey (10 adults) were at Hart Mine Marsh (figure 11). Palo Verde Ecological Reserve Phase 5 had the highest mean number of eggs detected per survey (three eggs) (figure 12). Caterpillars were not detected at a level useful for analyses (16 detected in 4 years) and are not included in these results.

Figure 10.—Total surveys conducted per site, 2015–19.

24 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Figure 11.—Mean adult sootywings detected per survey, 2015–19.

Figure 12.—Mean eggs detected per quailbush surveyed, 2015–19.

25 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Predictors of Sootywing Presence

Scatter plots of ranked data (figure 13) and Spearman’s rank correlations (table 2) indicate at most only weak associations of habitat characteristics with sootywing egg and adult counts. Egg abundance was positively correlated with adult abundance (rs = 0.35), leaf density (rs = 0.22), quailbush size (rs = 0.13 for area; rs = 0.12 for height and width), and leaf damage (rs = 0.15) (table 2). Adult abundance was positively correlated with quailbush area (rs = 0.41) as well as with height (rs = 0.39), width (rs = 0.0.40), and soil moisture (rs = 0.26) (table 2).

Figure 13.—Scatter plot matrix of ranked quailbush attributes and counts of eggs and adults with density histograms and linear fits for data collected at conservation areas, 2015–19. n = 267, except for pairs with leaf density or leaf size (n = 181) or adults (n = 82).

26 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Table 2.—Spearman’s rank correlation coefficients (ρ) of quailbush attributes and counts of sootywing eggs and adults (Asterisks indicate coefficients statistically different from 0 based on a t-test [uncorrected for multiple comparisons]. n = 267, except for pairs including leaf density or leaf size [n = 181] or adults [n = 82]).

Leaf Leaf Soil Parameter Adults Eggs density size Damage moisture % dry Area Width Height 0.39*** 0.12* -0.06 0.002 -0.02 0.17** -0.01 0.94*** 0.77***

Width 0.40*** 0.12* -0.008 0.01 -0.03 0.15* 0.002 0.94*** Area 0.41*** 0.13* -0.04 0.01 -0.03 0.17** -0.01 % dry 0.05 -0.02 -0.34*** -0.22** 0.08 -0.13* Soil moisture 0.26* 0.09 0.03 0.06 -0.03 Damage 0.07 0.15* 0.11 0.14 Leaf size 0.07 0.11 0.16* Leaf density 0.08 0.22** Eggs 0.35** * p < 0.05. ** p < 0.01. *** p < 0.001.

Both sets of stepwise logistic regression selected plant size (height or area) and leaf density as the only important predictors of sootywing presence. The best height model had a lower AICc (209.03) than the best area model (211.37). The Hosmer-Lemeshow goodness of fit test, which compares observed and expected values, indicated a good fit for the height model (p = 0.61) but not for the area model (p = 0.036). Model performance, reflected as the percentage of accurately predicted values by bootstrapping, was very similar: 68 and 69% values accurately predicted in models with area or height, respectively. Thus, the model with plant height and leaf density was chosen as the final, best model (table 3).

Leaf density had the largest effect on the probability of egg presence. Eggs were 6.7 times more likely to be on quailbush with high versus low leaf density, but the large CIs reflects relatively large uncertainty of this estimate (1.99, 22.24). The effect of medium leaf density was not statistically different from that of low density, as the 95% CI included 1. Effects of leaf density on the probability scale are shown in figure 14a. The effect size of plant height was smaller; a 1-foot increase in height increases the odds of eggs by only 16% (table 3). This is equivalent to ≈2 times greater likelihood of eggs for a 5-foot increase in plant area. On the probability scale, a doubling of plant height from 5 to 10 feet approximately doubles the probability of eggs from 12 to almost 22% (figure 14b).

27 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

Table 3.—Odds ratios for predictors from the best logistic regression model on egg presence (SE and 95% CIs are based on robust SE. p-values are for the Wald test of null hypothesis that odds ratios equal 1. n = 181.)

Odds Parameter ratio SE 95% CI z p Height 1.16 0.07 [1.03, 1.31] 2.44 0.015 Leafden [Medium] 2.08 1.09 [0.74, 5.79] 1.4 0.163 Leafden [High] 6.65 4.1 [1.99, 22.24] 3.07 0.002

Figure 14.—Predicted probability (with 95% confidence band or interval) of egg presence versus leaf density(left) and quailbush height (feet) (right).

DISCUSSION

Quailbush leaf density, height, and area were most strongly correlated with egg counts. The correlation between egg counts and area has been shown in previous research on host plant associations and sootywing egg presence, which found a 104% increased probability of egg presence per 0.2 square meters increase in plant area (Weisenborn and Pratt 2008). Quailbush height performed slightly better than area as a predictor when combined with leaf density in a logistic regression model, which suggests that height is a sufficient measurement for plant size in future monitoring. Model results indicate sootywing adults select larger

28 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

plants with denser leaf growth for oviposition. The finding does not point to specific management options other than maintaining conditions that support large plants with high leaf density, which would most likely be watering regime, but may also be associated with other factors not analyzed in this report, such as within-site quailbush dispersion and soil salinity. Leaf growth, percent dry quailbush, soil moisture, and caterpillar damage were not predictive of Sootywing presence. As these measurements do not currently provide value for measuring the quality of sootywing habitat, it may be beneficial to remove these measurements and focus surveys on the variables most predictive of sootywing presence in order to better inform management of habitat suitability.

Contrary to the soil moisture readings taken during surveys, measurements from long-term soil moisture sensors were not positively correlated with sootywing counts. This is most likely due to the very small sample size, as only four surveys had soil moisture data for the sites and times of data collection. Future surveys could focus on quailbush habitat with nearby soil moisture sensors in order to more clearly define the relationship between soil moisture and preferential quailbush attributes. Collecting more data near soil moisture data loggers may reveal a stronger relationship between soil moisture and egg presence. In the face of limited water rights per conservation area and an ongoing drought, a clearly defined level of water required for suitable sootywing habitat would be useful information for management.

Many of the variables taken were categorical visual estimates. In order to produce more accurate results when analyzing leaf density response to treatments, and sootywing presence relative to leaf density, a more quantitative method could be used. The LCR MSCP has access to a lidar database of restoration sites from 2016 to present. As lidar measures structure, it may be worthwhile to see whether the dataset can be used to identify sites with large, dense quailbush and to compare egg presence or adult abundance with those quailbush characteristics. If this is possible, it would reduce the amount of on-the-ground surveying necessary to provide an estimate of habitat suitability across creditable sootywing acreage. If it turns out that soil moisture is strongly correlated with the determinates of egg presence detected by lidar, this data may be useful to inform what watering regimes are used to maintain suitable sootywing habitat in land cover creditable for the species.

29 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

LITERATURE CITED

Burnham, K.P. and D.R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach, 2nd edition. New York: Springer-Verlag.

LCR MSCP (see Lower Colorado River Multi-Species Conservation Program).

Lower Colorado Multi-Species Conservation Program. 2015. LCR MSCP Mobile Electronic Field Form Guidebook – Sootywing. Bureau of Reclamation, Boulder City, Nevada.

Lüdecke D. 2020. sjPlot: Data Visualization for Statistics in Social Science. R package version 2.8.5. https://CRAN.R-project.org/package=sjPlot

Lüdecke D., D. Makowski, P. Wahhoner, and I. Patil. 2020. Assessment of regression models performance. R package version 0.5.0. https://easystats.github.io/performance/

Makowski, D., M.S. Ben-Shachar, I .Patil, and D. Lüdecke. 2019. Methods and Algorithms for Correlation Analysis in R. Journal of Open Source Software5(51), 2306. 10.21105/joss.02306.

Mazerolle, M.J. 2020. AICcmodavg: Model selection and multimodel inference based on (Q)AIC(c). R package version 2.3-1. https://cran.r-project.org/package=AICcmodavg/

Nelson, S.M., R. Wydoski, and S. Pucherelli. 2015. Monitoring of the MacNeill’s Sootywing and its Habitats, 2015 Annual Report. Bureau of Reclamation report submitted to the Lower Colorado River Multi-Species Conservation Program, Boulder City, Nevada.

Pratt, G.F. and W.D. Wiesenborn. 2009. MacNeill’s sootywing (Hesperopsis gracielae) (: Hesperiidae) behaviors observed along transects. Proceedings of the Entomological Society of Washington 111(3):698–707.

_____. 2011. Geographic distribution of MacNeill’s sootywing (Hesperopsis gracielae) (Lepidoptera: Hesperiidae) along the lower Colorado River floodplain. Proceedings of the Entomological Society of Washington 113(1):31–41.

31 Monitoring of the MacNeill’s Sootywing Skipper and its Habitats 2020 Annual Report

R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/

Venables, W.N. and B.D. Ripley. 2002. Random and mixed effects. Modern Applied Statistics with S. Springer, New York 271–300.

Wiesenborn, W.D. 2017. Monitoring MacNeill’s Sootywings at the Cibola Valley Conservation Area and Palo Verde Ecological Reserve, 2012 Annual Report. Bureau of Reclamation, Boulder City, Nevada.

Wiesenborn, W.D. and G.F. Pratt. 2008. Selection of Atriplex lentiformis host plants by Hesperopsis gracielae (Lepidoptera: Hesperiidae). Florida Entomologist 91:192–197.

_____. 2010. Visitation of heliotrope and western purslane flowers by Hesperopsis gracielae (Lepidoptera: Hesperiidae).

32

ATTACHMENT 1

Survey Area Maps

Attachment 1 – Page 1 Attachment 1 – Survey Area Maps

Attachment 1 – Page 2 Attachment 1 – Survey Area Maps

Attachment 1 – Page 3 Attachment 1 – Survey Area Maps

Attachment 1 – Page 4 Attachment 1 – Survey Area Maps

Attachment 1 – Page 5 Attachment 1 – Survey Area Maps

Attachment 1 – Page 6 Attachment 1 – Survey Area Maps

Attachment 1 – Page 7 Attachment 1 – Survey Area Maps

Attachment 1 – Page 8 Attachment 1 – Survey Area Maps

Attachment 1 – Page 9