The latest five-year review for this species (USFWS 2016) focused on habitat loss to urban development and OHV damage as the primary threats. It emphasized habitat protection from disturbance as the principal management objective. The only biotic threat identified as imminent was loss of native pollinators. Invasive species such as annual bromes were still considered a potential future threat of unknown impact, and herbivory and disease were not considered threats. In 2018, we initiated census and monitoring activities across four dwarf bear poppy populations utilizing UAV (drone) imagery analysis (Rominger and Meyer 2019, Rominger et al. in press). This followed a major dwarf bear poppy recruitment event in spring 2017. While we have not yet quantified all the data sets, we already have strong evidence of major differences among populations in both mortality rates and reproductive output. Observations both from the imagery and on the ground implicate rodent herbivory, predispersal seed predation by insects, and pathogen-caused disease as factors in increased mortality and reduced reproductive success at one population (White Dome). At the same time, we have observed an explosive increase in annual bromes at this population. There is a critical need to determine the level of these negative impacts across multiple populations, and whether the impacts occur only in association with massive increases in annual bromes in adjacent areas. We also need to know whether impacts are weather-related and therefore likely transient or whether they represent longer-term population trends. Annual brome invasion into closely adjacent areas may be driving indirect negative impacts in poppy populations that occupy largely uninvaded sectors of the gypsum habitat. If this proves to be the case, management to prevent extinction will need to shift to elimination of annual brome infestations in the gypsum environment.

The work proposed here is a continuation of our ongoing studies of factors affecting dwarf bear poppy population dynamics. Our overarching hypothesis here is that increased threats from herbivory and disease within poppy-occupied habitat are associated with annual brome infestation in closely adjacent areas, and that the threats will therefore be greater in populations suffering high levels of brome invasion. The study will include re-evaluation, with new objectives in mind, of existing data and imagery (as well as data and imagery to be obtained in spring 2021). We will also carry out an additional year of more targeted imagery and data acquisition in spring 2022. This will provide us with four years of drone imagery and associated data. Our proposed one-year study is intended to meet the following objectives: 1) Create drone imagery-based maps of annual brome invasion into different habitats within the gypsum environment (to be completed summer 2021) at two populations with differing levels of brome invasion (White Dome and Beehive Dome), and determine which habitats within the gypsum environment are conducive to annual brome success at each population and which habitats are still contributing to their exclusion. 2) Combine maps of brome distribution with our previously developed census maps that show poppy distribution across the gypsum environment at each population; use these to identify the degree of overlap between brome habitat and poppy habitat, as well as to locate areas where brome infestations are in close proximity to poppy subpopulations. 3) Carry out preliminary field and laboratory studies to determine the causal agents of rodent herbivory, insect herbivory, predispersal seed predation, and pathogen-caused disease impacts at the whole plant and inflorescence levels, and quantify the impacts of these threats at the two populations. 4) Re-evaluate monitoring imagery and field data across years for both populations to quantify the spatiotemporal pattern and extent of rodent (i.e., pocket gopher and/or antelope ground squirrel) whole plant herbivory, and examine its relationship with proximity to brome infestations. 5) Synthesize results of studies through spring 2022 to evaluate the hypothesis that the observed increase in threat level from annual brome invasion, herbivory and disease is associated with an unusual pattern of inter-annual variation in winter-spring precipitation and is therefore likely to be transient. The alternative hypothesis is that this increase in threat level is part of a long-term trend that is of more pressing concern for dwarf bear poppy conservation.

As described in the five-year review for dwarf bear poppy (USFWS 2016), our study will be the first to provide quantitative data to evaluate Listing Factor C of the ESA, namely the impact of disease and/or predation (including herbivory) as threats to population viability. Our study also addresses the following recommendation in Section 6.1 of the five-year review (p. 26): "Regular monitoring of invasive species in poppy habitat is recommended, and research on the direct and indirect impact of invasive species to the species (i.e., dwarf bear poppy) is needed". Our studies will provide information that will potentially improve management for conservation of dwarf bear poppy across federal, state, tribal, and private lands. By documenting serious negative impacts, both direct and indirect, from annual brome invasion, we can alert managers to the need to take effective preemptive control measures while infestations are small.

The drone pilots who will carry out flights are fully licensed by FAA and will follow all regulations pertaining to the use of UAV technology for image acquisition. We have permission from the BLM, SITLA (Utah State School Trust Lands), and TNC to carry out drone flights over these land ownerships and to carry out concomitant on-the-ground studies. We will obtain permits from Utah DWR for live-trapping to characterize rodent communities and will follow all pertinent regulations regarding the handling of live-trapped animals. The US Fish and Wildlife Service has funded the current phase of this dwarf bear poppy population dynamics study and is fully supportive of our continuing research on this species. TNC has funded an initial study of the annual brome invasion at the White Dome Preserve and is in full support of our continued work there.

Objective 1. To map the distribution of annual brome cover at White Dome and Beehive Dome will require drone image capture over each census area in spring and ideally again in summer to determine brome cover through color differencing (NDVI) methods, which measure the cover of green vegetation. By carrying out image capture at two dates it is possible to separate exotic annual cover, which loses greenness in the summer, from perennial vegetation, which retains some greenness. A possible difficulty with applying this method in 2021 is a dry spring with little or no annual brome cover. If this proves to be the case we will develop a method for evaluating past brome cover (2020) using the color of the litter patches that remain in place. Drone imagery will be captured and processed as described in Rominger and Meyer (2019) and analysis will be carried out in ArcPro GIS software. Once we have maps of brome cover in each area, we will overlay these on our previously developed maps of soil surface color type at each area (Rominger and Meyer 2019) to determine the degree of brome invasion as a function of surface color type. Poppy relative abundance was highly correlated with this habitat index in census imagery from 2018. Objective 2. It will be a straightforward process in to overlay the brome cover maps at each population onto the poppy census maps developed from the 2018 census (Rominger and Meyer 2019) to determine overlap between poppy-occupied habitat and brome-occupied habitat as well as proximity of areas of poppy occupancy to brome infestations. Objective 3. On-the-ground studies will include a continuation of the annual monitoring protocol (described in Rominger et al. in press) in late spring 2022. This includes obtaining drone imagery of monitoring plots at each of seven populations and carrying out on-the-ground reproductive output evaluation at each population. This will provide data on recruitment, survivorship, growth, and reproductive output over the period since spring 2021, and will also provide the imagery needed to evaluate current-year rodent herbivory at the two principal study sites in comparison to herbivory impacts in previous years (see below). In addition, we will perform field studies to quantify disease and insect predation both on inflorescences and at the whole plant level at the two principal study sites. Preliminary work on pathogen identification and pathogenicity testing is already underway, and this spring (2021) we will focus on identifying the organisms responsible for inflorescence and fruit damage. This will position us for a more systematic approach to quantification of the damage due to these threats in 2022. We will also characterize the rodent community at each site and examine rodent abundance and spatial distribution in relation to brome infested habitat and poppy habitat with live trapping using mark recapture methods. In the laboratory, we will continue our studies of the fungal pathogens responsible for poppy foliar diseases that can cause whole plant mortality as well as massive fruit abortion. As we process fruit collections from the monitoring reproductive output study, we will identify any insect herbivores and predispersal insect seed predators we find and will also quantify the observed damage. Objective 4. Our working hypothesis on rodent herbivory is that areas closer to extensive brome infestations will have more rodent herbivory, both within a population and across populations with different levels of brome infestation. To quantify mortality from rodent herbivory, we will utilize drone imagery over multiple years in monitoring plots to attempt to determine the cause of mortality from one year to the next. We have already determined that it is relatively easy to detect mortality caused by pocket gophers, which burrow underneath the plants and pull them into their burrows by the roots, leaving an obvious hole that is visible in drone imagery. Ground squirrel-caused mortality has also been observed; in this case the plant is chewed off at the taproot from above and left on the surface, where it can remain visible for some time. Plants left standing in place after death and still visible can be assumed not killed by rodent herbivores. Plants that are missing would have to be recorded as mortality due to unknown causes, but even with this missing data we should be able to quantify relative levels of herbivory and possibly to detect patterns of herbivory as a function of distance to a brome infestation. These analyses will complement data obtained from the trapping study. Objective 5. By summer 2022 we will have completed all the fieldwork and image analysis and will be able to synthesize the results from five years of monitoring to examine patterns of mortality and reproductive output by population and year. We will use our more detailed data sets from 2022 to evaluate the alternative hypotheses that the new threats we have documented are either causing transient effects due to unusual interannual variation in precipitation or that they are likely to have more long term negative impacts on one or more dwarf bear poppy populations.

We have designed and implemented a monitoring plan within each population, with nine 0.6 ha plots that are flown and evaluated annually using drone and GIS technology according to a well-defined protocol that will be transferable to land managers and their contractors.

Dr. Michael Stevens at Utah Valley University will be the Principal Investigator on this project and will oversee all aspects of the work, while Project Manager Kody Rominger will be directly involved in the day-to-day conduct of the research. Dr. Susan Meyer and Dr. Tara Bishop at the US Forest Service Rocky Mountain Research Station will serve as Co-Principal Investigators. Our research on dwarf bear poppy to date has been carried out with cooperation and funding from the US Fish and Wildlife Service, the BLM Utah State Office, the BLM St. George Field Office, Utah Department of Natural Resources, the Washington County Habitat Conservation Plan, and the Nature Conservancy. Volunteers from Southern Utah University, The Nature Conservancy, and the Red Cliffs Preserve as well as other local volunteers have helped in various aspects of our work. We will continue to work with these partners as we proceed with the proposed project.

If this monitoring is carried out beyond the timeframe of our research as outlined above, there will be a wealth of information to use for evaluating changing population status over the longer time frames needed to distinguish between transient and long-term trends. We hope to be involved directly in this work for at least three more years. Brome infestation in previously uninvaded rare plant habitat is worrisome. Adequate evaluation using long term monitoring data sets obtained over large areas is necessary to understand its impacts.This goal would be difficult to achieve without drone technology.

07/01/2021

06/30/2022

2022

We used both remote sensing and on the ground evaluation to address the primary objectives of the project. Drone imagery and GIS tools were used to map the spatial pattern of brome abundance in the high production year (2020) at White Dome based on residual litter cover a year later in 2021, a year of very low annual brome production. We also obtained drone imagery to examine annual brome abundance at Beehive Dome in 2021. We then used four years of drone monitoring imagery (nine 0.6 ha permanent plots at Beehive Dome, and seven at White Dome) and on the ground reproductive output monitoring (2019-2022) to evaluate and compare patterns of recruitment, survival, growth, and reproductive output at the two study populations. We used both drone imagery and on the ground studies to examine whether the increased levels of annual brome invasion at White Dome in 2020 were related to increased levels of rodent herbivory, insect foliar and capsule herbivory, and fungal pathogen-caused disease in comparison to Beehive Dome, where brome abundance had remained low. Lastly, we carried out a pilot GIS analysis of LANDSAT imagery using the greenness differencing methodology for two contrasting brome production years at White Dome and the surrounding area to determine the feasibility of tracking annual brome abundance across years with imagery from the LANDSAT archive with 30 m pixel resolution.

Our first project objectives were to use drone imagery to map annual brome patterns of invasion at White Dome and Beehive Dome in 2020, a high-precipitation year that resulted in a large increase in brome cover at White Dome, and to use these maps along with surface cover habitat classification and poppy census maps to determine the amount of overlap between invaded habitat and core poppy habitat. At Beehive Dome there was too little litter cover to use this method, which was not surprising based on field observations of low levels of brome invasion there except along washes. Mapped annual brome cover over the 209 ha flight area on the east side of the White Dome Preserve based on litter from the 2020 high-production year was 6.0% overall. The brome cover was not uniformly distributed. It was much higher and more continuous in areas closer to the northern, northeastern, and southern preserve boundaries, areas with large infestations in adjacent areas, showing the potential importance of propagule pressure. The core of the preserve area was much less invaded, and the area inside the fenced research area in the southwest corner had annual brome cover of only 2.3%, almost three times lower than the flight area considered overall. When annual brome distribution was examined relative to the surface cover classification based on surface soil color, brome cover values were three times higher on the dark crust and light crust surface cover classes than on the white cover class, whereas dwarf bear poppy densities from the 2018 drone census were five times higher on the white cover class than on the dark or light crust cover classes. This shows that the white knoll top habitat that supports most of the poppies is much less invaded than habitats where poppy densities are low, both locally and at a landscape scale. A total of 87% of the poppies in the 2021 flight area were found within the fenced research area in the 2018 census. In spite of this apparent habitat non-overlap and the lower overall brome cover within the research area, a spatial analysis of these two distribution data sets determined that 85% of the poppy points were <10m from a brome point, even in the research area. The likely explanation for this is that annual bromes can colonize the areas beneath shrub canopies, even on knoll tops in the white surface cover class, creating a polka-dot-like spread of very small invasion foci within poppy-occupied habitat. The poppies are rarely if ever found under shrub canopies, resulting in <1% intersection between brome and poppy points, and <5% of poppy points were within 1 m of a brome point in the research area where the majority of the poppies occurred. Values were similar for the flight area as a whole. Another factor in the relatively close proximity of brome points to poppy points in spite of low overall brome cover is the fact that the surface cover classes can also occur in close proximity on the landscape scale, placing more invaded habitat immediately adjacent to uninvaded habitat that supports poppies. We used the four-year dwarf bear poppy drone monitoring data set to determine whether brome invasion had a deleterious effect on population dynamics at White Dome relative to Beehive Dome, which was minimally invaded. The population of adult poppies that was present on the 0.6 ha monitoring plots was similar (923 plants at White Dome and 931 plants at Beehive Dome, respectively) when monitoring plots were established in 2019. These were either reproductively mature two-year-old plants from the large 2017 recruitment event or adults still present from previous recruitment events. The plants are not long-lived, so that survival curves that showed steep linear decline at both populations from 2019 to 2022 were not surprising. These survival curves were similar for the two populations over this period, and survival was actually higher in 2022 at White Dome (13% of original plants) than at Beehive Dome (1.2%) These results suggest that annual brome invasion has not impacted adult dwarf bear poppy survival. In addition, adult poppy mortality rates were not related to inter-annual variation in precipitation patterns, even though there was a dramatic difference between wet and dry years during this period. Patterns of year-to-year change in plant diameter were also very similar at the two sites. Seed production in dwarf bear poppy is generally high, but reproductive output monitoring at the two populations over the four-year monitoring period revealed some between-population differences. In 2019 and 2020, fruit set was uniformly high at both locations, above 95%. In 2021, both populations had reduced fruit set, but the effect was much more pronounced at White Dome (53% vs. 92%). In 2022, fruit set was <50% at both populations. These lower fruit-set years coincided with dry winters that may have affected pollinator abundance or plant resource status. Filled seed number per capsule was higher at Beehive Dome than at White Dome in all but the final year of monitoring. This resulted in higher filled seed number per flower (fruits per flower x seeds per fruit) at Beehive Dome in the first three years. In spite of high fruit abortion in 2022, the White Dome plants were able to produce as many seeds per flower as Beehive Dome plants due to exceptionally high seed fill that year. These results indicate that the White Dome plants overall are likely to have lower seed production, but there is no evidence that the year of high brome abundance resulted in reduced reproductive output there, as that was the best year of the four for White Dome in terms of seed production per flower. We evaluated rodent activity-related mortality using drone imagery across four years at the two sites. The rodents responsible were surmised to be antelope ground squirrels (Ammospermophilus leucurus), diurnal herbivorous rodents commonly observed at both sites. We evaluated a 1 m circle centered on each point where a poppy was newly missing each year in all of the monitoring plots at each site. We scored each circle as 'no new disturbance', 'rodent hole' when a hole was found instead of the poppy that was present the previous year, or as 'rodent digging' when fresh mounds of soil were observed. Over the three-year period, there was somewhat higher rodent-associated mortality at Beehive Dome than at White Dome, but the major difference between the sites was in the timing of rodent-associated mortality, which was much higher at White Dome in 2020 and much higher at Beehive Dome in 2021. Neither site had significant rodent-associated mortality in 2022. The reason for this site difference in the year of peak rodent activity is not known, but there is no strong evidence that annual brome cover in 2020 was implicated in rodent-associated mortality there, especially as even higher mortality was observed the following year at Beehive Dome in the absence of significant annual brome cover. We also evaluated herbivory at the scale of foliar and capsule damage for plants included in the reproductive output study. Most of this damage was due to a variety of insects. Levels of herbivory were similar at the two sites but trended higher at Beehive Dome. The most striking pattern in the data set was very low levels of herbivory in the high rainfall years the first two years of the study, but much higher levels during the dry years, especially the final year of the study. There was no indication that associated high brome cover increased small-scale herbivory on poppies, which rarely caused major damage to the plants. We attempted to carry out evaluation of disease levels on poppy plants in the reproductive output study each year, but the accuracy was hampered by the timing of data collection, which may not have coincided with the time of maximum disease development. We determined through genetic ITS fingerprinting using cultures grown from infected tissue that the major pathogen causing both foliar and inflorescence damage was a species of the fungal genus Alternaria. Disease symptoms were present across the first three years at White Dome and present at levels as high as or higher than those observed at Beehive Dome each of those years. Beehive Dome only had significant disease incidence in 2020, and both populations exhibited a lack of disease symptoms in 2022. These results are hard to interpret, but there is no compelling case for an association with annual brome invasion at White Dome, as in the 2020 high brome production year, both populations had similarly high levels of disease (ca. 40% of plants with symptoms). There was little direct evidence that disease was a cause of mortality, but this could be because the timing of data collection did not permit us to fully make this connection. In summary, we did not obtain any evidence that the brome invasion at White Dome at its current level has substantially increased any indirect threats to dwarf bear poppy persistence. The population and individual level data from two populations with dramatically different levels of brome invasion showed largely similar patterns across four years of monitoring. However, annual brome invasion may still pose a threat in and of itself, in spite of the habitat segregation between annual bromes and poppies evident in the analysis. The presence of annual bromes under shrub clumps throughout the preserve has placed an annual brome seed source within occupied poppy habitat. This is a scenario that has unfolded in other places that initially seemed to have relatively low invasion potential. Given the right sequence of years, it could be possible for the annual bromes modify the surface at the local invasion edges sufficiently with their own litter to enable slow but inexorable spread into the interspaces between the shrubs. Once the fine fuel layer of brome standing litter becomes continuous, the risk of fire increases exponentially. This means that dealing with the annual brome problem before it escalates to this level should be a high priority in dwarf bear poppy habitat throughout its range. Please see the attached report document for a more complete description of our studies.

Based on this study as well as an earlier long-term monitoring study at a different population, the basic population dynamic pattern for dwarf bear poppy could be described as a 'boom and bust' cycle, with high recruitment success during favorable periods, that is, years with high winter-spring precipitation, followed by high seed production for a few years as well as steady mortality rates regardless of weather patterns. If the period between recruitment years is longer than the life span of most adult plants, densities can fall to very low levels. However, the presence of a large persistent seed bank enables population recovery in response to subsequent favorable years. This means that habitat once occupied is still potentially occupied and is likely to support adult poppies in the future. The causes of mortality vary with environmental conditions, but the overall pattern remains the same. Direct anthropogenic disturbance can disrupt this pattern, making protection from disturbance the key component in management of existing populations regardless of their current status in terms of density, as is the recommendation of the USFWS and also the current management practice. The apparently recent spread and encroachment of annual bromes into some gypsum badlands habitats that support dwarf bear poppies adds another layer of complexity to management. We propose examining these invasion patterns in more detail both historically from LANDSAT imagery and through drone-based mapping in years with more living brome cover that would permit use of greenness differencing or multispectral mapping methods. This would include evaluation of potential source populations that may be driving invasion into poppy habitat through increased propagule pressure. We also propose a continuation of our pilot studies on the feasibility of safely using pre-emergent herbicides to control annual bromes in the gypsum badlands habitat, that is, without any negative effects on dwarf bear poppy. We believe that the best approach to annual brome control would be to accurately map the infestations using drone technology, then use precision herbicide application with a spray drone that can apply the chemical to small, well-defined areas that are programmed into its flight plan. This technology is already available at relatively low cost for agricultural use. This would minimize surface disturbance, cost, and labor associated with on the ground herbicide application and would permit a level of control that would be impossible with traditional aerial herbicide application. We look forward to working with land managers across the multiple land ownerships that support dwarf bear poppy to explore this option.