New Mexico meadow jumping mouse Project Report for 2015-2017CEAP WWWG Agreement No.: 68-7482-15-505Reporting Period: May 2015 – December 2017Project Title: Occurrence of New Mexico meadow jumping mouse on the Apache-Sitgreaves NationalForestsPI: Carol L. Chambers, P.O. Box 15018, School of Forestry, Northern Arizona University, Flagstaff, AZ86011, E-mail: [email protected], Phone: 928-523-0014Executive Summary1. Between 2015 and 2017, we captured 110 New Mexico meadow jumping mice (Zapus hudsoniusluteus) in Arizona (Apache-Sitgreaves National Forest, private lands) and New Mexico (Santa FeNational Forest). From these captures, we obtained 86 fecal samples for genetic analysis of diet,63 hair samples for stable isotope analysis of diet, 77 buccal samples for future genetic analysis,and deployed 12 transmitters to determine home range and habitat use.2. We developed non-invasive track plate to detect New Mexico meadow jumping mice. Createdtraining video now adopted by USFWS to survey for the species in the Southwest. Publicationdescribing the process is in review.3. To predict habitat use, we surveyed 75 locations on the Apache-Sitgreaves National Forest in2016, collected habitat data, and developed an occupancy model predicting probability ofpresence. Identified 7 habitat elements that were positively associated with presence ofjumping mice: vegetation height, stream width, percent cover of alder (Alnus spp.), percent forbcover, presence of wild ungulates (elk, deer), soil moisture, and percent sedge cover. Identified1 habitat element that was negatively associated with presence of jumping mice: streamgradient. Livestock grazing was negatively associated with jumping mice but not a strongpredictor in this model.4. Identified new locations for the jumping mouse outside of designated critical habitat.5. Conducted a diet analysis using stable isotopes and metagenomics that identified jumping miceas herbivorous, feeding largely on C3 plants.6. After identifying diet as herbivorous from stable isotopes and metagenomics, we developed aninexpensive and rapid metabarcoding assay that genetically identifies plant diet for anyherbivorous species. We successfully tested it on jumping mouse, mule deer, and pronghorn.We determined that jumping mice have a diverse diet dominated by forbs and grasses thatdiversifies greatly in late summer and supports gain in mass prior to hibernation.7. Calculated home ranges using Minimum Convex Polygon and 95% Kernel Probability averaged1.21 ha and 1.05 ha, respectively. Animals remained close to streams and side drainages. Theymoved an average of 6 m from stream; maximum distance from stream averaged 15 m. Daynests for jumping mice were identified as grassy bolus structures near streambanks.8. For outreach, we have drafted a Wildlife Habitat Evaluation Guide, created a videodemonstrating the non-invasive track plate method, trained others in the field to use our trackplate method, gave 20 presentations on research results at meetings for resource managers andscientists, and are preparing manuscripts from our work. Peer-reviewed publications provide asound scientific basis for management decisions. We submitted a manuscript on the trackplating method to Wildlife Society Bulletin which has been reviewed favorably and is in revision.We anticipate 6 additional publications from this project.9. We acquired additional funding from Forest Service and Arizona Game and Fish Department tosurvey for and examine diet of New Mexico meadow jumping mice.

CEAP WWWG Agreement No.: 68-7482-15-505C. L. ChambersProject Report for 2015-2017Northern Arizona University10. Based on the work supported by NRCS on diet of the jumping mouse, our genetics ResearchSpecialist, Dan Sanchez, applied for and was awarded a prestigious National Science FoundationFellowship ( to work on his PhD studying diet and population geneticsof the New Mexico meadow jumping mouse. He was one of 2000 selected for this award of13,000 that applied.Research NeedsRadio telemetry in AZ and NM ( 50,000)Radio telemetry in CO ( 35,000)JustificationThe New Mexico meadow jumping mouse (Zapus hudsonius luteus) is considered a riparianobligate that uses tall, dense herbaceous vegetation along perennial flowing water such as streams,ditches, and wet meadows. Although jumping mice are found in riparian areas with moist soils, they alsouse adjacent dry upland areas beyond the floodplain to nest, bear and raise young, and hibernate.Jumping mice need high quality food sources prior to hibernation to accumulate fat reserves; seeds arethought to provide these reserves. The diet of the jumping mouse is not clearly defined but observationsfrom other subspecies and from a small sample in New Mexico indicated they may shift from adominance on insects shortly after emergence from hibernation in spring to seeds just before enteringhibernation in fall. Surveys identified 8 geographic areas in Arizona, New Mexico, and Colorado wherejumping mice occurred but some areas remain to be surveyed. Livestock grazing is thought to affectjumping mouse habitat. To develop recommendations for grazing (timing, frequency) and allowcontinuation of livestock grazing while promoting habitat for the jumping mouse, this study identifiesdistribution, habitat requirements, diet, and population parameters of the species. field surveys for jumping mice at both historic and new locations (including on privatelands wherever allowed), collect non-invasive genetic samples from jumping mice and habitatdata (vegetation and grazing data) at each study site.Estimate probabilities of detection and occupancy for the jumping mouse and relate this tohabitat data. Use these occupancy models to predict the relative probability of occupancy acrossthe study area and map potential habitat on private lands.Determine population relatedness and diet using genetic approaches from jumping mice (liveand museum specimens) captured in Arizona and New Mexico. Determine patch size andconnectivity needed to support jumping mice.Develop effective grazing and restoration approaches for the jumping mouse.Overview2015We used 2015 as a pilot year to trap sites where New Mexico meadow jumping mice werepreviously captured, develop a non-invasive approach to detecting jumping mice (track plating), andcollect fecal and hair samples that we analyzed for diet. Our work was focused along streams that hadbeen previously sampled on the Apache-Sitgreaves National Forests in Arizona (Frey 2011, Hicks 2014).We used these sites to test capture methods, test track plate designs, and increase the probability ofcapturing jumping mice for collecting fecal samples for dietary analysis. However, we also live-trappedand track plated at new locations within proposed critical habitat (Figure 1).2

CEAP WWWG Agreement No.: 68-7482-15-505C. L. ChambersProject Report for 2015-2017Northern Arizona UniversityIn May, we visited sites to assess the habitat quality of potential habitat for the jumping mouse.Information gathered from this trip along with geospatial data was used to guide site selection fortrapping during the summer field season. We hired a summer field crew (5 technicians) to capture smallmammals and 2 professional botanists (Judith Springer, NAU, and Kirstin Phillips, Museum of NorthernArizona) to provide detailed lists of plant species present in areas where we captured jumping mice.Field crew personnel were trained in small mammal capture and identification techniques andvegetation sampling. During the training period, the field crew also worked on the development of atrack plate design and a small mammal track reference guide that could be used in the field foridentification of jumping mouse tracks. We collected field data from 15 June to 18 September 2015. Weused 5 weeks for live-capture methods and 5 weeks for track plate methods to sample jumping mice.Vegetation sampling was conducted at all trap sites where jumping mice were captured. All data wereentered and summarized. On 10 November we presented initial results from our field season at theApache-Sitgreaves National Forest Supervisors Office in Springerville.2016In summer 2016, we compared our non-invasive track plate method to live trapping at 14 sitesand found similar rates of detection for jumping mice. We surveyed 75 sites (n 59 using track plating, n 16 using live trapping) on the Apache-Sitgreaves National Forests in Arizona (Figure 2) and developedan occupancy model to predict presence of jumping mice. At 12 sites where we detected jumping micethere were livestock grazing allotments with varying levels of grazing. We surveyed 10 sites on the SantaFe National Forest in New Mexico (Figure 3) to identify sites where jumping mice were present. Jumpingmice were captured or detected with track plates at sites outside designated critical habitat. Wecollected hair samples for stable isotope analysis of diet, fecal samples for genetic analysis of diet, andbuccal samples for future population genetic analysis.We hired a crew leader, field crew (5 technicians) to capture small mammals, a field crew (4technicians) to survey vegetation and habitat structure, and a botanist to provide detailed lists of plantspecies present in areas that we surveyed for the jumping mouse. Field crew personnel were trained insmall mammal capture and identification techniques and vegetation sampling. We surveyed for jumpingmice and collected vegetation data from June to September 2016. Vegetation sampling was conductedat all sites where we surveyed for jumping mice, regardless of whether we captured the species. All datawere entered and summarized. On 18 November, we presented initial results from our field season atthe Apache-Sitgreaves National Forest Supervisors Office in Springerville. On 12 December, wepresented results, including our occupancy model, to the Forest Service biologists meeting inAlbuquerque.2017Between June and September 2017, we surveyed 17 locations on the Apache-SitgreavesNational Forests (Figure 4), 11 sites on the Santa Fe National Forest (Figure 5), and 10 sites on theLincoln National Forest (Figure 6) for presence of jumping mice. Vegetation sampling was conducted atall sites where we surveyed for jumping mice, regardless of whether we captured the species. Weidentified occupied sites outside designated critical habitat. We collected hair samples for stable isotopeanalysis of diet, fecal samples for genetic analysis of diet, and buccal samples for future populationgenetic analysis. We conducted a radio telemetry pilot study to identify home ranges (Minimum ConvexPolygon [MCP] and Kernel Density Probability [KD; 95%]) and distances moved from riparian areas onthe Apache-Sitgreaves and Santa Fe National Forests. We collected vegetation data in home rangesusing the cover board approach. We identified day nests constructed of grass-like materials.For field work conducted from June to September 2017, we hired a crew leader, 4 technicians tocapture small mammals, 2 technicians to survey vegetation and habitat structure, and a botanist to3

CEAP WWWG Agreement No.: 68-7482-15-505C. L. ChambersProject Report for 2015-2017Northern Arizona Universitydocument plant species present in areas that we surveyed for the jumping mouse. We trained field crewpersonnel in small mammal capture and identification techniques and vegetation sampling. Data fortrack plating, trapping, habitat, and telemetry were entered and summarized.We submitted for publication a manuscript on the track plating procedure. This paper is inreview with Wildlife Society Bulletin. We trained an independent (Arizona Game and Fish Department)crew in June on track plating methods for jumping mice that they subsequently used on the ApacheSitgreaves National Forests. On 27 November, we presented our results to the Forest Service Wildlife,Fish, and Rare Plants Session in Albuquerque.MethodsSite selectionIn 2016, to develop our occupancy model, we used a Geographic Information System (ArcGIS,ESRI, Redlands, California) to select 100 potential sampling sites on the Apache-Sitgreaves NationalForests based on our criteria for elevation ( 2740 m), stream classification (perennial vs intermittent),riparian vegetation (based on RMAP and PNVT data), and distance to roads. We set a maximum distancefrom roads of 2 km (1.2 miles) for feasibility of access by field crews. We then randomly selected sitesthat met all of the above criteria and stratified these sites based on grazing and recreational use. Weused 5 vegetation categories from RMAP riparian vegetation (Arizona alder, Herbaceous riparian,willow/thin leaf alder, Upper montane conifer/Willow, Ponderosa pine/willow). We used 4 riparianvegetation categories from PNVT (Cottonwood-Willow Riparian Forest, Mixed Broadleaf deciduousriparian forest, Montane Willow riparian forest, wetland/cienega riparian areas).For other sites and years, sites were provided by National Forest personnel. In Arizona, wesurveyed land in private ownership after the landowner requested surveys.TrappingWe live trapped small mammals using non-folding Sherman traps (LNA 7.6 x 8.9 x 22.9 cm). Wetrapped sites in June, July, or August each year. Traps were set for 1 to 4 days; we opened traps eachevening from 3 hours prior to dusk ( 17:00) and dusk ( 20:00), checked them each morning within 4hours of dawn ( 0500 to 0900), and closed traps during the day. During 2015, we set 40 traps along thestream channel on the bank (adjacent to water), 40 traps in the ecotone, and 40 traps in the upland todetermine whether jumping mice moved beyond riparian areas. Traps were placed 3 to 5 m apart. For2016 and 2017, we used a trapping transect at each site that consisted of 80 traps placed 3 to 5 m apartonly along the stream channel on the bank where we best detected jumping mice in 2015. In 2015, webaited traps with sweet feed. We shifted to a mixture of steel-cut oats and peanuts for subsequent yearsas this provided a simpler bait to detect and remove during genetic analysis of diet.For each individual captured using Sherman traps, we recorded trap number, species captured,sex, age, and reproductive condition unless high capture rates affected our ability to check traps in atimely manner. In these cases we only collected species or genus for non-target species. When jumpingmice were captured, we collected all morphological information as for other species but also measuredtotal body length, tail length, hind foot, ear length, and mass. We also collected fecal, hair, and buccal(cheek swab) samples from individual jumping mice for diet and DNA analysis. If we expected to radiocollar a jumping mouse, we limited data collection to sex and mass to avoid additional stress to animals.Track PlatingWe used track plates that we developed to survey jumping mice. We trapped sites in betweenJune and September each year. Track plates consisted of a plastic box with 2 entry holes cut into sides, apiece of self-adhesive, clear paper cut to the dimensions of the enclosure and attached sticky side upwith double sided tape, and a piece of felt fabric, saturated in solution of mineral oil and carpenter’s4

CEAP WWWG Agreement No.: 68-7482-15-505C. L. ChambersProject Report for 2015-2017Northern Arizona Universitychalk. Roofing felt provided a cover that prevented precipitation from entering the enclosure. Trackplates were placed 3 to 5 m apart along riparian areas (same method as with Sherman live traps). Weplaced bait along the top or edge of the felt pad of the enclosure to attract rodents to the boxes. Trackplates were checked for 2 to 4 days. Details of this method are presented in the track plate trainingvideo we created: i2x0Ydc1XVM.Vegetation and HabitatPlant richness and dominanceVegetation data were collected at sites that were track plated or trapped. Richness anddominance of plants were collected between June and October. A botanist surveyed riparian, ecotone,and upland areas along each trap site and identified plant species present in the survey area. Eachsurvey area was 250 x 500 m and encompassed the riparian area that included the trap transect as wellas the ecotone and upland on either side of the riparian area. Dominance used the Daubenmire CoverClass categories (0 0%, 1 0 – 5% [midpoint 2.5%], 2 5 – 25% [15%], 3 25 – 50% [37.5%], 4 50 –75% [62.5%], 5 75 – 95% [85%], 6 95 – 100% [97.5%]).StructureWe collected structure and habitat data which consisted of stream characteristics (e.g., gradient,width), elevation, aspect, vegetation height (total and laid over), Robel pole measurements, grazing sign,recreation, and beaver sign. Vegetation data were also collected at the microhabitat level usingDaubenmire plots at 34% of the trap locations at each site (40 of 120 traps per site in 2015, 27 of 80traps per site in 2016 and 2017). We collected vegetation data at all sites where we captured NewMexico meadow jumping mice.Following development of our occupancy model in 2016, we reduced the number of variablescollected but added sign of recreation. For 2016 and 2017, we used logistic regression in a preliminaryanalysis to compare vegetation and habitat features between sites with and without presence ofjumping mice.Diet AnalysisWe used 3 techniques to analyze diet of jumping mice: isotope analysis, metagenomics, andmetabarcoding. Isotope analysis was conducted on hair samples to determine long term trends in dietbecause carbon and nitrogen stable-isotope ratios (δ13C and δ15N) from assimilated foods areincorporated into tissues (e.g., hair) of the consumer. In general, δ15N is enriched 3 to 5 per mil fromproducer to consumer and thus indicate trophic feeding level. Diet can also be described based oncarbon-isotopic distinction between C3 (dicots, cool-season grasses) or C4 (warm-season grasses) plants.Using metagenomics we identified species (e.g., plants, vertebrates, invertebrates, bacterialpathogens) present in the feces and the dominance of each in the fecal sample. This approach indicatedthe focus of metabarcoding should identify plant species present in feces. We continue to barcodeplants found in the study areas and believed to be diet items that are not yet in the Barcode of LifeDatabase (BOLD; Ratnasingham and Hebert 2007). By barcoding new species, we are better able toidentify diet items to species. Presently we are confident in our identification to genus and for someplants, to species.For analysis of fecal DNA, we collected feces from traps in which we captured a jumping mouseusing the sample preservation and DNA extraction methods of Walker et al. (2016). In brief, we PCRamplified a plant-specific ITS2 region (Chen et al. 2010) in fecal DNA of jumping mice and recorded thesequence mixtures in parallel using an Illumina MiSeq. We then classified taxonomy for each sequenceusing a naïve Bayes classifier (RDP) that we trained on a global, plant DNA database (Barcode of LifeDatabase).5

CEAP WWWG Agreement No.: 68-7482-15-505C. L. ChambersProject Report for 2015-2017Northern Arizona UniversityTelemetryWe radio collared 12 jumping mice (10 males, 2 females) in July and August on 2 national forests(Apache-Sitgreaves and Santa Fe). To limit handling stress, only sex of the animal was taken although forseveral animals we were also able to take mass. We used 0.47 g BD-2XC transmitters (Holohil SystemsLtd, Ontario, Canada) fitted with a 25 to 27 mm TYGON sleeve to cover the antenna wire and preventabrasion to the neck of the animal. We used the antenna wrapped through the collar to attach to theanimal. Total mass of the BD-2XC collar including sleeve and crimp was 0.6 g. To minimize effects to theanimal we limited mass of the collar to 5% of its body mass (i.e., a 0.6 g collar should be placed on ananimal 12 g).We attempted 1 observation per day and 3 observations per night per animal. Weapproached each a