Progress 02/25/24 to 02/24/25
Outputs
Target Audience:The target audience for the Beescape and BeeWinterWise toolsincludesbeekeepers, gardeners, conservationist, landscape architects, growers as well as students. The target audience for the workflows and data sets that we generated and which are housed in the BeeShiny toolare the scientific community. This year, we engaged over 20,000 stakeholders through the online tool and through presentations to stakeholder groups and in classrooms. We had more than 75 people participate in the Ecospatial Summit, respresenting 34 universities, government agencies, and non-profit groups, where individual provided feedback on the BeeShiny tool and learned to use it to obtain data on land use, habitat quality, and weather conditions at user-selected sites across the United States.? Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?During this reporting period, this project has supported two graduate students, Lily Houtman, PhD, Geography, Penn State and Anais Ostroski, PhD Environmental Engineering, University of Pittsburgh, one postdoctoral scholar, Dr. Jaya Sravanthi Mokkapati, one undergraduate student, Hunter Reed, Penn State and one part-time research assistant, Garrett Sisk, Penn State. The project also supported Dr Heather Grab, Senior Lecturer at Cornell and, since summer2024, Assistant Professor, Entomology, Penn State and David McLaughlin, Data Visualization Software Engineer, Institute for Computational and Data Sciences, Penn State. Several other students and early career researchers have been involved in the project, but supported financially through different sources, including Dr. Gabriela Quinlan, NSF Postdoctoral Research Fellow in Biology, Entomology, Penn State and subsequently USDA-ARS postdoctoral fellow and Dr. Kevin Li, USDA SCINet Postdoctoral Fellow. All individuals have received advanced training in geospatial data analysis, ecological modeling, human-centered data visualization, cartography, decision support tool development and assessment, as well as ecology, entomology and agriculture.? How have the results been disseminated to communities of interest? We presented information on Beescape at the following conferences and university seminars between March 2024 and February 2025: MultiTroph Symposium, China, Scialog Neurobiology and Changing Ecosystems Symposium, USDA Pollinator State of the Science Workshop, EcoSpatial Summit, Kent State University Ben Foote Memorial Lecture in Ecology. We presented information on Beescape at the following stakeholder meetings: Pennsylvania State Beekeepers Association. Worchester MA Beekeeping Club, Penn State Extension webinar series, University of Florida Two Bees in a Podcast, Honey Love Beekeeping Group, Penn State Ag Progress Days, Pennsylvania Conservation Districts Annual Conference, Penn State Millennium Cafe. Additionally, Beescape is used as an assignment in Grozinger's general education course, ENT 222, Honey Bees and Humans, which has 70 students from across Penn State University. The assignment helps students gain an understanding of how local and landscape features influence habitat quality for bees. What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
Objective 1. Improve usability and accuracy of the seasonal forage quality index, through (A) stakeholder assessment workshop, and (B) improving the thematic, temporal, and spatial resolution of the data sources. The team developed and published on approaches to increase both thematic and temporal resolution. Merging the USDA Cropland Data Layer (CDL) with the National Vegetation Classification layer provided by LANDFIRE greatly increased the number of vegetation classes, and additionally provided a way to characterize species abundance by vegetation class and region. The merged raster maps were published (Kammerer et al., 2022), and the associated scientific manuscript has been published (Kammerer et al., 2024). Weexamined the weather parameters and habitat factors that influenced wild bee abundance and diversity. Early-season bee communities responded primarily to landscape resources, including flowering trees and wetland habitats. Mid to late-season communities were more influenced by local conditions, though bee diversity was negatively impacted when sites were embedded in highly agricultural landscapes. Soil composition had complex impacts, and likely reflects effects on plant community flowering. This study was published in Landscape Ecology in 2024. Using the PA State Beekeepers annual winter loss survey (>3000 responses from 2017-2022), we demonstrated that the number of different types management interventions to control parasites was positively correlated with survival. Random forest analysis highlighted the importance of summer and fall temperature and precipitation and summer forage on winter survival, suggesting that foraging resources during the growing season is essential for winter survival. Colonies managed using multiple treatments were resilient to a broader range of weather conditions. This manuscript was published in the Journal of Insect Science in 2024. Objective 2. (A) Develop a national map of economic value of pollination services at the 30-m resolution scale. (B) Identify "hot spots" where economic value of pollination service is threatened due to increased pollination demand and risks and decreasing resources for pollinators. The team leveraged fine scale bee visitation and production data to map pollinator foraging and corresponding yields at high spatial and temporal resolution in wild blueberries. We evaluated the contribution of field topography, microclimate and crop bloom intensity on both wild and managed bees using machine learning and structural equation models. Topographic variation resulted in microclimates with differences of as much as 10°C and 29% relative humidity. Honey bees were the predominant visitors and were correlated with wild bee visitation which did not vary with distance from honey bee hives. Both groups had substantial temporal and microclimatic overlap in their foraging profiles, though wild bees foraged earlier and at a wider range of conditions. The positive effect of warmer microclimates on wild bee foraging was similar in magnitude to the negative effect of forest edge distance. Flower density, which was greater in sites with warmer microclimates, was the primary driver of foraging for both wild bees and honey bees as well as yields. After accounting for flower density, no relationship was observed between either wild or honey bee visitation rates and blueberry yields, suggesting that pollination services were not limiting yield. Our findings indicate that microclimates within fields can have stronger effects on crop yield than other well recognized factors, mediated by their effects on floral density and bee foraging. This study is in review at Agriculture, Ecosystems and Environment. The team worked on Bayesian inference models to understand the relationship between wild bee abundance, soybean yields, and toxicity from pesticides. Using previously downscaled information, we discuss the approach adopted by the land sharing/sparing discourse to inform land use intensity and its relationship with biodiversity. Results suggest that production practices, proxied by pesticide use, could be used as indicators to determine land-use intensity decision-making rather than production only. The manuscript describing these studies is in preparation for submission to a journal. Objective 3. Create a next-generation Beescape (Beescape NexGen) that optimizes the ability of diverse stakeholders to understand, explore, use, and contribute to information on spatiotemporal variation in pollinator health and ecosystem services. In summer 2024, we conducted another web-based usability study using the redesigned Beescape (launched in 2023) with 43 beekeepers (Houtman et al., in review). Assessment included training exercises, three analytical tasks, usability and utility evaluation rating questions, and a series of prompts focused on ideas for future improvements. Usability results were overall quite high, and showed considerable improvement over the same assessment of our original prototype. In addition to the new features added in 2023, we modified Beescape to work well on mobile devices andwe developed pipeline for adding data on plant species found in specific locations and habitats, using information from the Biota of North America program. We hope to integrate the data from BONAP into Beescape for Pennsylvania as a proof of principle by May 2025. Additionally, in 2024, we launched the BeeWinterWise tool, which allows beekeepers and other to view winter colony survival data at the county level in Pennsylvania from 2017 onwards, and examine the impacts if different pest management strategies (https://pollinators.psu.edu/bee-health/bee-winterwise ) Travis Flohr (Assistant Professor of Landscape Architecture, Penn State University)worked with an undergraduate student(Hunter Reed) to select 25 pollinator-attractive plants used in urban gardens and developed scores of resiliency to varying habitat and weather conditions. They developed fact sheets for these plants to share with gardeners and the public. Objective 4. Foster a collaborative scientific community to catalyze research in the field of bee health and landscape ecology. We developed open-source code and repositories to allow others to readily access our data and methodologies to assess the role of land use, weather and climate impacts on bee health. For the published studies, data and code have been provided as part of the supplementary materials. In Fall 2024, we launched the BeeShiny platform and associated github repository (https://pollinators.psu.edu/research/beeshiny). Beeshiny is a web-based tool that makes it easier to access the land use, habitat quality, weather and climate data currently featured in Beescape for selected locations (including county and state scale) and allow users to upload and generate data for multiple locations simultaneously. Additionally, users can readily obtain information for previous years, which is not currently a feature available in Beescape. In October 2024, we hosted the Ecospatial Summit (ecospatialsummit.com), where we invited individuals who had used Beescape or data in the Beescape layers to test and provide feedback on the new BeeShiny tool. More than 75 people participated, representing 34 universities, government agencies, and nonprofit groups.
Publications
- Type:
Other Journal Articles
Status:
Under Review
Year Published:
2025
Citation:
Sisk, G., Ostroski, A, Dillard, T., Hall, B., Goslee, S.C., Grozinger, C.M., Khanna, V., and H. Grab. "Unraveling microclimate effects on pollinator foraging and crop yield in lowbush blueberry" Agriculture, Ecosystems and Environment.
- Type:
Other Journal Articles
Status:
Under Review
Year Published:
2025
Citation:
Houtman, L., Robinson, A.C., McLaughlin, D., and C.M. Grozinger. "Evaluating the Usability and Utility of a Spatial Decision Support System for Pollinator Ecology" Ecological Informatics.
|
Progress 02/25/21 to 02/24/25
Outputs
Target Audience:The target audience for the Beescape and BeeWinterWise tools includes beekeepers, gardeners, conservationist, landscape architects, growers as well as students. The target audience for the workflows and data sets that we generated and which are housed in the BeeShiny tool are the scientific community. We engaged over 20,000 stakeholders per yearthrough the online tool and through presentations to stakeholder groups and in classrooms. In Fall 2024, had more than 75 people participate in the Ecospatial Summit, respresenting 34 universities, government agencies, and non-profit groups, where individuals provided feedback on the BeeShiny tool and learned to use it to obtain data on land use, habitat quality, and weather conditions at user-selected sites across the United States. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Over the lifetime of the project, the project directly supported four graduate students (Prestby, Unger, Gray, Houtman, Ostroski), two undergraduates (Egan, Reed) one postdoctoral fellow (Mokkapati), one research assistant (Sisk), and two research scientists (Grab, McLaughlin). Several early career researchers were involved in the project but received salary support from other sources (Kammerer, Quinlan, Caradima, Li). All individuals received advanced training in geospatial data analysis, ecological modeling, human-centered data visualization, cartography, decision support tool development and assessment, as well as ecology, entomology and agriculture. Additionally, Beescape is used as an assignment in Grozinger's general education course, ENT 222, Honey Bees and Humans, which has an annual enrollment of 70 students. The assignment helps students gain an understanding of how local and landscape features influence habitat quality for bees, and how different bee species have different seasonal and nesting needs.? How have the results been disseminated to communities of interest?The results of the project have been presented at 40 scientific talks to universities, government agencies, and scientific conferences. A selection of conferences where this work was presented is listed here: Pollination Conservation and Ecology Webinar Series from the US FWS National Conservation Training Center,Entomological Society of America Annual Meeting, Foundation for Food and Agricultural Science Annual Board Meeting, North American Pollinator Protection Campaign annual meeting, International Association for Landscape Ecology North American Annual Meeting, USDA ARS SCINet and AI COE Fellows Conference, Ecological Society of America Annual Meeting, American Geophysical Union, EuroCarto conference, International Conference on Pollinator Biology, Health and Policy, USDA/EPA Pollinator State of the Science Workshop; British Ecological Society Annual Meeting; International Cartographic Association Meeting. The results have been presented to 12 stakeholder conferences. A selection is highlighted here: the Pennsylvania State Beekeepers Association. State College, PA; Southern Adirondack Beekeeper Association; Penn State Master Gardeners, Bristol County Beekeepers. Montgomery County Beekeepers Association, Worchester MA Beekeeping ClubPennsylvania GreenGov Webinar Series, Pennsylvania Conservation Districts Annual Conference. Additionally, Beescape has over 20,000 unique users annually. What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
Objective 1. Improve usability and accuracy of the seasonal forage quality index, through (A) stakeholder assessment workshop, and (B) improving the thematic, temporal, and spatial resolution of the data sources. The team developed approaches to increase both thematic and temporal resolution by merging the USDA Cropland Data Layer (CDL) with the National Vegetation Classification layer provided by LANDFIRE. The merged raster maps were published (Kammerer et al., 2022), and the associated scientific manuscript has been published (Kammerer et al., 2024). The team examined the role of trees in urban areas in supporting bee populations, using the urban tree map for Philadelphia, PA, and data provided by beekeepers. We demonstrated that tree cover was positively correlated with honey bee colony growth. We found increasing thematic resolution of land use categories was more important than increased spatial resolution. The team developed a workflow to use citizen science data on distribution and flowering times of plant species to produce regional floral abundance curves. The ability to track both weather and flowering patterns will make it possible to offer a near-real-time assessment of floral resources, and perhaps even a forecast. The team developed a machine-learning workflow to downscale gridded temperature data to provide cumulative growing degree day and precipitation deficit estimates at the 30 m resolution on annual and daily timesteps to improve the spatial resolution of modeling for both plant and pollinator distribution and seasonality. We demonstrated that the weather parameters and habitat factors influenced spring and summer wild bee abundance and diversity. This study was published in Landscape Ecology in 2024. Using the PA State Beekeepers annual winter loss survey (>3000 responses from 2017-2022), we demonstrated that the number of different types management interventions to control parasites was positively correlated with survival. Colonies managed using multiple treatments were resilient to a broader range of weather conditions. This manuscript was published in the Journal of Insect Science in 2024. We analyzed the spatial and temporal drivers of honey production at the state-wide level in the continental United States. Ecoregion, and presumably the associated plant communities, is the major driver of overall honey yield, and this is modulated by land use and weather. Her data also demonstrated the importance of soil type, which likely influences plant growth and productivity. This paper was published in Environmental Research Letters in 2023. Objective 2. (A) Develop a national map of economic value of pollination services at the 30-m resolution scale. (B) Identify "hot spots" where economic value of pollination service is threatened due to increased pollination demand and risks and decreasing resources for pollinators. The team downscaled our previously completed economic value of pollination services analysis from county level to 30m resolution. This economic layer was integrated into Beescape NexGen in summer 2023. The team leveraged fine scale bee visitation and production data to map pollinator foraging and corresponding yields at high spatial and temporal resolution in wild blueberries. Our findings indicate that microclimates within fields can have stronger effects on crop yield than other well recognized factors, mediated by their effects on floral density and bee foraging. This study is in review at Agriculture, Ecosystems and Environment. The team worked on Bayesian inference models to understand the relationship between wild bee abundance, soybean yields, and toxicity from pesticides. The manuscript describing these studies is in preparation for submission to a journal. Objective 3. Create a next-generation Beescape (Beescape NexGen) that optimizes the ability of diverse stakeholders to understand, explore, use, and contribute to information on spatiotemporal variation in pollinator health and ecosystem services. Through structured surveys and workshops, we evaluated how Beescape addressed the needs and goals of different stakeholders (beekeepers, growers, conservationists, scientists) (Prestby et al., 2023; Robinson et al., 2021). The platform was redesigned and launched as Beescape NexGen in 2023. In summer 2024, we conducted another web-based usability study with 43 beekeepers (Houtman et al., in review). Usability results were overall quite high, and showed considerable improvement over the same assessment of our original prototype. The new platform has the following features compared to the previous version: A new landing page for Beescape, which contains an instruction video and more detailed information. More flexibility in selecting the size and shape of the landscape areas the user wants to assess. More detailed information about the types of land use and habitats in the selected areas. New information on the economic value of pollination services to crops Data on monthly average temperature and total precipitation for current year, previous year, and 10-year normals Information on pollinators and flowering plants that are currently present in the landscape, which can help users know which plants will provide floral resources in their area during specific months. The ability to print out a report card for selected sites Modifications to allow for viewing on mobile devices Additionally, we developed pipeline for adding data on plant species found in specific locations and habitats, using information from the Biota of North America program. We hope to integrate the data from BONAP into Beescape for Pennsylvania as a proof of principle by May 2025. In 2024, we launched the BeeWinterWise tool, which allows beekeepers and other to view winter colony survival data at the county level in Pennsylvania from 2017 onwards, and examine the impacts if different pest management strategies (https://pollinators.psu.edu/bee-health/bee-winterwise ) We developed a database of ecoregional plant species that support pollinator populations in Pennsylvania and surrounding states. Weselected 25 pollinator-attractive plants used in urban gardens and developed fact sheets for these plants to share with gardeners and the public. Objective 4. Foster a collaborative scientific community to catalyze research in the field of bee health and landscape ecology. We developed open-source code and repositories to allow others to readily access our data and methodologies to assess the role of land use, weather and climate impacts on bee health. For the published studies, data and code have been provided as part of the supplementary materials. In Fall 2024, we launched the BeeShiny platform and associated github repository (https://pollinators.psu.edu/research/beeshiny). Beeshiny is a web-based tool that makes it easier to access the land use, habitat quality, weather and climate data currently featured in Beescape for selected locations (including county and state scale) and allow users to upload and generate data for multiple locations simultaneously. Additionally, users can readily obtain information for previous years, which is not currently a feature available in Beescape. In October 2024, we hosted the Ecospatial Summit (ecospatialsummit.com), where we invited individuals who had used Beescape or data in the Beescape layers to test and provide feedback on the new BeeShiny tool. More than 75 people participated, representing 34 universities, government agencies, and nonprofit groups.
Publications
|
Progress 02/25/23 to 02/24/24
Outputs
Target Audience:The Beescape tool was accessed by 22,000 unique users over this period. The target audience includes beekeepers, growers, conservationists, and members of the public. Beescape has been used by researchers to evaluate sites for their studies, and several researchers contacted us to request a tool that would allow them to obtain habitat quality scores and climatic information for multiple sites; this tool is currently in development. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?During this reporting period, this project has supported three graduate students, Darcy Gray, MSc, Ecology (graduated in December 2024 and now a Remote Sensing Scientist at the One Acre Fund in Kenya), Penn State, Lily Houtman, PhD, Geography, Penn State and Anais Ostroski, PhD Environmental Engineering, University of Pittsburgh and one undergraduate student, Garrett Sisk, Penn State. The project also supported Dr Heather Grab, Senior Lecturer at Cornell and soon Assistant Professor, Entomology, Penn State and David McLaughlin, Data Visualization Software Engineer, Institute for Computational and Data Sciences, Penn State. Several other students and early career researchers have been involved in the project, but supported financially through different sources, including Dr. Gabriela Quinlan, NSF Postdoctoral Research Fellow in Biology, Entomology, Penn State and Dr. Kevin Li, USDA SCINet Postdoctoral Fellow. All individuals have received advanced training in geospatial data analysis, ecological modeling, human-centered data visualization, cartography, decision support tool development and assessment, as well as ecology, entomology and agriculture. How have the results been disseminated to communities of interest?Grozinger presented information on Beescape at the following conferences and university seminars between January and December 2023: International Conference on Pollinator Biology, Health and Policy, USDA/EPA Pollinator State of the Science Workshop; Penn State Communication Arts and Sciences 2023 Summer Symposium; Penn State Institute for Computational and Data Science Annual Symposium; University of California, Santa Barbara; University of Muenster. Grozinger presented information on Beescape at the following stakeholder meetings: Pennsylvania State Beekeepers Association. Bristol County Beekeepers. Montgomery County Beekeepers Association. Pennsylvania GreenGov Webinar Series. Penn State College of Agricultural Sciences College Connections. NW PA Beekeepers Association Field Day. Lancaster Beekeeping Club. Penn State Science Policy Society Goslee presented information on Beescape at the International Conference on Pollinator Biology, Health and Policy, Northeast Pasture Consortium stakeholder meeting. Houtman and Robinson presented information on Beescape at the International Cartographic Association Meeting. Gray presented information on Beescape to the PA State Beekeepers Association, at the International Conference on Pollinator Biology, Health and Policy and the the International Conference on Landscape Ecology. What do you plan to do during the next reporting period to accomplish the goals?Objective 1 During the next reporting period, we will survey beekeepers on their experience with using Beescape NexGen and gather information on recommendations and suggestions. The downscaled weather data will be used to improve the phenological modeling and mapping and the newly-published integrated land cover data to produce more nuanced maps and predictions of pollinator floral resources. Objective 2 We will update the economic data in Beescape NextGen using the 2022 Census of Agriculture data. We will finalize our analysis of the effects of microclimate on bee foraging and pollination services in blueberry fields. We will finalize the workflow for generating the fine scale spatial and temporal economic valuation and publish our results. In the next reporting period, further statistical analysis involving the relative efficiencies of spatial units will be carried for the soybean and almond analysis. The team will also analyze economic value of pollination services at 30m resolution in conjunction with crop productivity assessment to gain insights on vulnerable areas and assess trends in agricultural efficiency of pollination dependent crops. Objective 3 We will modify the Beescape platform to allow the tool to be used easily used on mobile devices. Objective 4 The team will develop a R-Shiny app that will be allow PA beekeepers to review winter survival data. We will create an R Shiny interface that allows user to update multiple locations and obtain data on land use, weather and climate, and economic data. We will offer a multi-day symposium and workshop for researchers to learn how to integrate our approaches and models into their own research, and discuss future research and applications related to the Beescape platform.
Impacts
What was accomplished under these goals?
Objective 1. Improve usability and accuracy of the seasonal forage quality index, through (A) stakeholder assessment workshop, and (B) improving the thematic, temporal, and spatial resolution of the data sources. Graduate student Lily Houtman recruited 50+ beekeepers to participate in a new survey (for spring 2024) to evaluate the newly designed Beescape NexGen platform. The team has focused on both thematic and temporal resolution during the reporting period. Merging the USDA Cropland Data Layer (CDL) with the National Vegetation Classification layer provided by LANDFIRE greatly increased the number of vegetation classes, and additionally provided a way to characterize species abundance by vegetation class and region. The merged raster maps were published (Kammerer et al., 2022), and the associated scientific manuscript has been published (Kammerer et al., 2024). The team developed a workflow to use citizen science data on distribution and flowering times of plant species important for pollinators to produce regional floral abundance curves, both long-term and for specific years. The ability to track both weather and flowering patterns will make it possible to offer a near-real-time assessment of floral resources, and perhaps even a forecast. The team developed a machine-learning workflow to downscale gridded temperature data to provide cumulative growing degree day and precipitation deficit estimates at the 30 m resolution on annual and daily timesteps to improve the spatial resolution of modeling for both plant and pollinator distribution and seasonality. Former graduate student Melanie Kammerer and current postdoctoral scholar Kevin Li examined the weather parameters and habitat factors that influenced wild bee abundance and diversity. Early-season bee communities responded primarily to landscape resources, including flowering trees and wetland habitats. Mid to late-season communities were more influenced by local conditions, though bee diversity was negatively impacted when sites were embedded in highly agricultural landscapes. Soil composition had complex impacts, and likely reflects effects on plant community flowering. This study is in review in Landscape Ecology. Using the PA State Beekeepers annual winter loss survey (>3000 responses from 2017-2022), graduate student Darcy Gray demonstrated that the number of different types management interventions to control parasites was positively correlated with survival. Random forest analysis highlighted the importance of summer and fall temperature and precipitation and summer forage on winter survival, suggesting that foraging resources during the growing season is essential for winter survival. Colonies managed using multiple treatments were resilient to a broader range of weather conditions. Gray and Penn State Extension Educator Robyn Underwood developed a streamlined and more user-friendly survey of winter survival and beekeeper management and launched this in May 2023. The number of responses with GIS locations increased significantly. Postdoctoral researcher Gabriela Quinlan spearheaded an analysis of the spatial and temporal drivers of honey production at the state-wide level in the continental United States. Ecoregion, and presumably the associated plant communities, is the major driver of overall honey yield, and this is modulated by land use and weather. Her data also demonstrated the importance of soil type, which likely influences plant growth and productivity. This paper was published in Environmental Research Letters in 2023. Objective 2. (A) Develop a national map of economic value of pollination services at the 30-m resolution scale. (B) Identify "hot spots" where economic value of pollination service is threatened due to increased pollination demand and risks and decreasing resources for pollinators. The team downscaled our previously completed economic value of pollination services analysis from county level to 30m resolution. This economic layer was integrated into Beescape NexGen in summer 2023. With the 2022 Census of Agriculture data now available, this will be updated to reflect these more recent yield and economic information. The team leveraged fine scale bee visitation and production data to map pollinator foraging and corresponding yields at high spatial and temporal resolution in wild blueberries. We evaluated the contribution of field topography, microclimate and crop bloom intensity on both wild and managed bees using machine learning and structural equation models. Warmer microclimates generated by topographic variation in the field drive both crop flowering intensity and wild bee foraging. Improved understanding of the drivers of fine scale pollinator foraging will allow for higher resolution estimates of both risk to bees as well as their contributions to crop production. The team worked on bayesian inference models to understand the relationship between wild bee abundance, soybean yields, and toxicity from pesticides. Using previously downscaled information, we discuss the approach adopted by the land sharing/sparing discourse to inform land use intensity and its relationship with biodiversity. Preliminary results suggest that production practices, proxied by pesticide use, could be used as indicators to determine land-use intensity decision-making rather than production only. Objective 3. Create a next-generation Beescape (Beescape NexGen) that optimizes the ability of diverse stakeholders to understand, explore, use, and contribute to information on spatiotemporal variation in pollinator health and ecosystem services. Beescape NexGen was launched in May 2023, and has received several additional updates since then. The new platform has the following features compared to the previous version: More flexibility in selecting the size and shape of the landscape areas the user wants to assess. More detailed information about the types of land use and habitats in the selected areas. New information on the economic value of pollination services to crops Data on monthly average temperature and total precipitation for current year, previous year, and 10-year normals Information on pollinators and flowering plants that are currently present in the landscape, which can help users know which plants will provide floral resources in their area during specific months. The ability to print out a report card for selected sites The Center for Pollinator Research created a new landing page for Beescape, which contains an instruction video and more detailed information. Information about the new Beescape platform was shared with stakeholders through a new story (https://pollinators.psu.edu/news/new-updates-come-to-beescape-an-online-tool-for-supporting-pollinators) and was featured in USDA's Science in Seconds series (@USDAScience Twitter:https://twitter.com/USDAScience/status/1692559655401029668). In the last year, Beescape has been visited by more than 22,000 unique users. Objective 4. Foster a collaborative scientific community to catalyze research in the field of bee health and landscape ecology. We are developing open-source code and repositories to allow others to readily access our data and methodologies to assess the role of land use, weather and climate impacts on bee health. The methods described in Objective 1 have been used as the basis of training materials on geospatial workflows for high performance computing. Data and methodologies are being published as stand-alone products where appropriate. An R package and associated documentation facilitating access to climate and weather data that supports other objectives has been developed. While some portions of this package are tailored to the specific server and datasets used in this project, the R package can be extended to support other workflows, and to allow collaborators effective access to these data.
Publications
- Type:
Journal Articles
Status:
Under Review
Year Published:
2024
Citation:
Kammerer, M. Iverson, A.L., Li, K., Tooker, J.F., and C.M. Grozinger. Seasonal bee communities vary in their responses to resources at local and landscape scales: implication for land managers Landscape Ecology (in review).
- Type:
Journal Articles
Status:
Accepted
Year Published:
2024
Citation:
Gray, D., Goslee, S., Kammerer, M, and C.M. Grozinger. Effective pest management approaches can mitigate honey bee colony winter loss (Apis mellifera) across a range of weather conditions in small-scale, stationary apiaries . Journal of Insect Science, in press.
- Type:
Journal Articles
Status:
Published
Year Published:
2024
Citation:
Kammerer, M., Iverson, A.L., Li, K, and S. C. Goslee. Not just crop or forest: an integrated land cover map for agricultural and natural areas. Sci Data 11, 137 (2024). https://doi.org/10.1038/s41597-024-02979-w
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Quinlan, G. M., Miller, D.A.W., and C.M. Grozinger. Examining spatial and temporal drivers of pollinator nutritional resources: Evidence from five decades of honey bee colony productivity data Environmental Research Letters 18(11): 114018 DOI 10.1088/1748-9326/acff0c (2023).
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Quinlan, Gabriela M., and Christina M. Grozinger. "Honey bee nutritional ecology: From physiology to landscapes." Advances in Insect Physiology https://doi.org/10.1016/bs.aiip.2023.01.003 (2023).
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Robinson, Anthony C., Timothy J. Prestby, Dave McLaughlin, Lily Houtman, Patrick Dudas, and Christina Grozinger. "Geovisualization for Pollinator Decision Support with Beescape NexGen." Abstracts of the ICA 6 (2023): 212.
|
Progress 02/25/22 to 02/24/23
Outputs
Target Audience:Our target audience includes beekeepers, growers, conservationists, researchers, policymakers, and members of the public (particularly individuals who have land that they manage or garden).? Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?This project has supported three graduate students, Timothy Prestby, PhD, Geography, Penn State, Lily Houtman, PhD, Geography, Penn State and Anais Ostroski, PhD Environmental Engineering, University of Pittsburgh and one undergraduate student, Maia Egan, Landscape Architecture, Penn State. Several other students and early career researchers have been involved in the project, but supported financially through different sources, including Dr. Melanie Kammerer, USDA SCINet Postdoctoral Fellow, Dr. Gabriela Quinlan, postdoctoral fellow, Entomology, Penn State, Dr. Heather Grab, postdoctoral fellow, Penn State, Dr. Kevin Li, USDA SCINet Postdoctoral Fellow, Darcy Gray, MSc graduate student and Fulbright Fellow, Penn State and Don Hagedorn, MGIS student, Penn State. All individuals have received advanced training in geospatial data analysis, ecological modeling, human-centered data visualization, cartography, decision support tool development and assessment, as well as ecology, entomology and agriculture. How have the results been disseminated to communities of interest? Kammerer presented updates on forage index at the seminar series of the USDA ARS Post-doc Network (online, March 2022), and USDA ARS Pollinator Brown Bag seminars (online, March 2022). Kammerer presented updates on floral landscape resources at the International Association for Landscape Ecology North American Meeting (April, 2022), and the Ecological Society of America Annual Meeting (August, 2022). Grozinger presented information on Beescape at the following invited university seminars between January and December 2022: University of Freiburg, FRIAS Natural & Life Science Colloquium ,Wright State University, Carmichael Lecture in the Environmental Sciences; Central State University, College of Engineering, Science, Technology and Agriculture; Montana State University, Department of Plant Sciences and Plant Pathology; Penn State University, Bortree Lecture, Department of Veterinary and Biomedical Sciences Grozinger presented information on Beescape at the following stakeholder meetings: Pennsylvania State Beekeepers Association. State College, PA. November 2022; Southern Adirondack Beekeeper Association. Online. October 2022; Penn State Master Gardeners. A Gardener's Guide to Change Symposium. Virtual. February 2022. Ostroski presented the findings of the teams work at the American Geophysical Union (AGU) Fall meeting in Chicago IL in December 2022. Anais Ostroski, Christina Grozinger, and Vikas Khanna. Coupling Earth Observations, Machine Learning, and Optimization to Uncover Productivity Drivers in Insect Pollination-Dependent Crops, Poster presentation at the AGU Fall Meeting, Chicago IL, December 14, 2022. Robinson presented findings from user evaluation work at the 2022 EuroCarto conference in Vienna, Austria in September, 2022. What do you plan to do during the next reporting period to accomplish the goals? Objective 1 During the next reporting period, the team will finalize the improved land cover map, develop strategies for updating it as appropriate, and create a workflow linking the phenology data with the land cover information. The team will publish papers related to how land use, weather and management influence winter bee survival. The team will develop a streamlined and more user-friendly survey for the Pennsylvania State Beekeepers to generate annual data on winter losses that will be more readily incorporated into our statistical models. Strong cross-linkages between Objectives 1 and 3 will ensure that these results are incorporated into Beescape in useful and appropriate ways. Objective 2 We will finalize the workflow for generating the fine scale spatial and temporal economic valuation and publish our results. In the next reporting period, further statistical analysis involving the relative efficiencies of spatial units will be carried for the soybean and almond analysis. The team will also analyze economic value of pollination services at 30m resolution in conjunction with crop productivity assessment to gain insights on vulnerable areas and assess trends in agricultural efficiency of pollination dependent crops. Objective 3 We will complete the new interface for Beescape NexGen by March 2023, and collect stakeholder feedback in the summer. We will develop a user-video for Beescape that will be dissminated through Penn State Extension. We will develop report cards for users to obtain information on potential improvements to habitat quality, including suggestions of pollinator-friendly plant species. We will seek to incorporate data from automated hive scales placed under honey bee colonies in partnership with Broodminder, Inc. We will also incorporate temporal data into Beescape. Objective 4 The land cover map and workflow will be published and made public. We will develop plans for a series of workshops for researchers to learn how to integrate our approaches and models into their own research.
Impacts
What was accomplished under these goals?
Objective 1. Improve usability and accuracy of the seasonal forage quality index, through (A) stakeholder assessment workshop, and (B) improving the thematic, temporal, and spatial resolution of the data sources. The team has focused on both thematic and temporal resolution during the reporting period. Merging the USDA Cropland Data Layer (CDL) with the National Vegetation Classification layer provided by LANDFIRE greatly increased the number of vegetation classes, and additionally provided a way to characterize species abundance by vegetation class and region. The merged raster maps were published (Kammerer et al., 2022), and the associated scientific manuscript is in review (Kammerer et al., 2023). The team developed a workflow to use citizen science data on distribution and flowering times of species important for pollinators to develop regional floral abundance curves, both long-term and for specific years. The ability to track both weather and flowering patterns will make it possible to offer a near-real-time assessment of floral resources, and perhaps even a forecast. The team examined the role of trees in urban areas in supporting bee populations. Using the urban tree map for Pittsburgh, PA, the team created a density map of tree-based resources for bees in urban Pittsburgh. This was integrated with information about registered apiaries in Pittsburgh, to identify sections that would benefit from the incorporation of additional floral resources for bees. Graduate student Darcy Gray demonstrated that tree cover in Philadelphia, PA, was positively correlated with honey bee colony growth. She found increasing thematic resolution of land use categories was more important than increased spatial resolution. Using the PA State Beekeepers annual winter loss survey,Gray demonstrated that the number of different types management interventions to control parasites was positively correlated with survival outcomes. Additionally, detailed analysis of seasonal weather conditions highlighted the importance of summer and fall temperature and precipitation on winter survival, suggesting that preparation for winter is a sensitive period in a honey bee colony life cycle. Team member Gabriela Quinlan spearheaded an analysis of the spatial and temporal drivers of honey production at the state-wide level in the continental United States. Her data demonstrated that ecoregion, and presumably the associated plant communities, is a major driver of overall honey yield, and this is modulated by land useand weather. Her data also demonstrated the importance of soil type, which likely influences plant growth and productivity. Objective 2. (A) Develop a national map of economic value of pollination services at the 30-m resolution scale. (B) Identify "hot spots" where economic value of pollination service is threatened due to increased pollination demand and risks and decreasing resources for pollinators. The team downscaled our previously completed economic value of pollination services analysis from county level to 30m resolution.This resulted in the production of economic value of pollination services base maps at the field scale for the years 2012 and 2017 (agricultural census years where uncertainty is lowest and dataset is most complete) which will be integrated into Beescape NexGen. During the reporting period, the team worked on a framework to integrate machine learning and an optimization-based method to evaluate agricultural production systems for their relative efficiency in resource management focusing on insect-pollination dependent crops. There is a gap in the literature regarding the inclusion of ecosystem services, such as pollination, in productivity assessments. We focused on soybeans and almonds due to their importance to US agriculture and diverging levels of pollinator dependencies. We utilized USDA-NASS agricultural census information, satellite observations from the MODIS sensor, and climate data from PRISM to downscale and estimate crop yield at 250m scale. These datasets were employed in machine learning models. For soybeans, non-parametric models such as random forest and k-nearest neighbors were shown to perform the best. For almonds, principal component regression with Bayesian lasso was employed due to high dimensionality. The downscaled yield was then coupled with Data Envelopment Analysis to understand how biotic and abiotic resources are used and what drives relative efficiency in crop production. The machine learning models were able to predict yield values for both crops with high accuracy, with coefficient of determination of 0.87 for soybeans and 0.93 for almonds. Preliminary results suggest that excess pesticide use drives inefficiency in soybeans, and less than 2% of 250x250m pixels had the maximum relative efficiency score. These areas are referred to as "benchmarks" that transform inputs into outputs more efficiently. These findings were presented at the American Geophysical Union (AGU) Fall meeting in December 2022 held in Chicago (IL). Objective 3. Create a next-generation Beescape (Beescape NexGen) that optimizes the ability of diverse stakeholders to understand, explore, use, and contribute to information on spatiotemporal variation in pollinator health and ecosystem services. In 2021, we conducted a one-day workshop to collect focus group data and paper prototypes for potential Beescape NexGen interfaces with a group of 21 stakeholders. A detailed report synthesizing the results of these knowledge elicitation activities was published (Prestby et al 2022). The project team developed a new graphical mockup including interaction details for Beescape NexGen. Beescape NexGen will allows userto select areas of interest using radial and polygon shapes to retrieve key pollinator habitat and health indicators. Habitat quality indicesare presented using a color scale, and users can obtain information on percentages of different habitats found in their selected areas. Users can view recent crowdsourced bee and plant observations from iNaturalist to determine which wild bee species are active and which plant species are currently blooming. Local climate patterns and year-to-date climate observations are available. The interface is in its final stages of development and should be available for the public by March 2023. Additionally, new team member Travis Flohr (Assistant Professor of Landscape Architecture, Penn State University) has been developing a database of ecoregional plant species that support pollinator populations in Pennsylvania and surrounding states. Travis worked with Maia Egan (Undergraduate Student of Landscape Architecture) to conduct a survey of published brochures, websites, databases, books, and articles documenting plant floral resources for pollinators in Pennsylvania and surrounding states. Using this information, Travis created a database structure, data dictionary, and protocol to enter plants into a database. Maia and Travis entered a pilot test of 80 plants with 66 plant attributes. Objective 4. Foster a collaborative scientific community to catalyze research in the field of bee health and landscape ecology. We are developing open-source code and repositories to allow others to readily access our data and methodologies to assess the role of land use, weather and climate impacts on bee health. The methods described in Objective 1 have been used as the basis of training materials on geospatial workflows for high performance computing. Data and methodologies are being published as stand-alone products where appropriate. An R package and associated documentation facilitating access to climate and weather data that supports other objectives has been developed. While some portions of this package are tailored to the specific server and datasets used in this project, the R package can be extended to support other workflows, and to allow collaborators effective access to these data.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Prestby, T.J., Robinson, A.C., McLaughlin, D., Dudas, P.M., and C.M. Grozinger. Characterizing user needs for Beescape: A spatial decision support tool focused on pollinator health Journal of Environmental Management 325: 116416 (2023). https://doi.org/10.1016/j.jenvman.2022.116416
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Robinson, A.C., Prestby, T.J., McLaughlin, D., Dudas, P., and Grozinger, C.M. (2022) Evaluating User Needs for Geovisualization of Pollinator Health and Ecosystems. European Cartographic Conference (EuroCarto), Vienna, Austria, September 19-21.
|
Progress 02/25/21 to 02/24/22
Outputs
Target Audience:Our target audience includes beekeepers, growers, conservationists, researchers, policymakers, and members of the public (particularly individuals who have land that they manage or garden). Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?This project has supported two graduate students, Timothy Prestby, PhD, Geography, Penn State and Mason Unger, MS, Environmental Engineering, University of Pittsburgh. Several other students and early career researchers have been involved in the project, but supported financially through different sources, including Dr. Melanie Kammerer, USDA SCINet Postdoctoral Fellow, Dr. Gabriela Quinlan, postdoctoral fellow, Entomology, Penn State, Dr. Bogdan Caradima, postdoctoral fellow, Penn State, and Darcy Gray, MSc graduate student and Fulbright Fellow, Penn State. All individuals have received advanced training in geospatial data analysis, ecological modeling, human-centered data visualization, cartography, decision support tool development and assessment, as well as ecology, entomology and agriculture. How have the results been disseminated to communities of interest? Grozinger has presented this project as an invited speaker at the Penn State Sustainability Institute virtual webinar series (online, February 2021), Pollination Conservation and Ecology Webinar Series from the US FWS National Conservation Training Center (online, June 2021), a symposium at Entomological Society of America Annual Meeting (Denver, CO and online October 2021), Foundation for Food and Agricultural Science Annual Board Meeting (Online, October 2021) and the North American Pollinator Protection Campaign annual meeting (online, October 2021). Goslee presented the statistical methodology being used to model plant and insect abundance distributions at the International Association for Landscape Ecology North American Annual Meeting (online, April 2021). Kammerer presented updates on forage index at the Penn State Sustainability Institute virtual webinar series (online, March 2021), workshop of the USDA ARS SCINet pollinator working group (online, August 2021), USDA ARS SCINet and AI COE Fellows Conference (online, November 2021), seminar series of the USDA ARS Post-doc Network (online, March 2022), and USDA ARS Pollinator Brown Bag seminars (online, March 2022). We also organized a two day Center for Pollinator Research Symposium in State College, which had 90 participants that spanned multiple departments and colleges within Penn State, as well as from industry, government, and conservation agencies. The work of the Beescape team was prominently featured in talks and posters. Additionally, we held a one day workshop following the symposium, which included beekeepers, growers, conservationists and researchers to evaluate and discuss the Beescape decision support tool (see Objective 3 discussion above). What do you plan to do during the next reporting period to accomplish the goals? Objective 1 During the next reporting period, the team will finalize the improved land cover map, develop strategies for updating it as appropriate, and create a workflow linking the phenology data with the land cover information. Strong cross-linkages between Objectives 1 and 3 will ensure that these results are incorporated into Beescape in useful and appropriate ways. Objective 2 We will finalize the workflow for generating the fine scale spatial and temporal economic valuation and publish our results. Objective 3 We will complete the new interface for Beescape NexGen by July 2022, and collect stakeholder feedback in the fall. We plan to submit a manuscript describing the results of our first stakeholder workshop a swell., Objective 4 The land cover map and workflow will be published and made public.
Impacts
What was accomplished under these goals?
Objective 1. Improve usability and accuracy of the seasonal forage quality index, through (A) stakeholder assessment workshop, and (B) improving the thematic, temporal, and spatial resolution of the data sources. The stakeholder assessment workshop described in detail under Objective 3 identified several themes around spatial and temporal scale that would make Beescape more useful for both large and small apiaries, particularly a more nuanced assessment of the utility of surrounding land use for pollinators, and the ability to track floral resource availability over the course of the growing season while comparing it to previous years or long-term means. The team has focused on both thematic and temporal resolution during the reporting period. The USDA Cropland Data Layer (CDL) conveys detail on crops grown each year, but delineates only broad categories of non-agricultural vegetation. Merging the CDL with the National Vegetation Classification layer provided by LANDFIRE greatly increased the number of vegetation classes, and additionally provided a way to characterize species abundance by vegetation class and region. The team developed a workflow to use citizen science data on distribution and flowering times of species important for pollinators to develop regional floral abundance curves, both long-term and for specific years. Data on floral abundance by species collected by colleagues can now be used to develop a comprehensive map of floral resource availability across space and time. The ability to track both weather and flowering patterns will make it possible to offer a near-real-time assessment of floral resources, and perhaps even a forecast. Objective 2. (A) Develop a national map of economic value of pollination services at the 30-m resolution scale. (B) Identify "hot spots" where economic value of pollination service is threatened due to increased pollination demand and risks and decreasing resources for pollinators. During the reporting period, the team downscaled our previously completed economic value of pollination services analysis from county level to 30m resolution. This was accomplished by leveraging spatial data from the USDA census and survey on pollination dependent crop acreage, yield, and price paired with crop dependence values from literature. In order to preserve the temporal aspects in this study, missing acreage values were reconciled by using higher level reported data (state, national) in the specific year of analysis. The spatial data set was then mapped according to FAO classifications to the cropland data layer to match the cell size and alignment of the seasonal forage, wild bee nesting, and insecticide indices. This resulted in the production of economic value of pollination services base maps at the field scale for the years 2012 and 2017 (agricultural census years where uncertainty is lowest and dataset is most complete) which will be integrated into Beescape NexGen. Users will be able to analyze crop systems for their economic value which could assist in colony placement, motivating conservation, and other localized decision making. The creation of this fine scale spatial and temporal economic valuation and integration with other models will provide insights on the security and vulnerability of pollination services nationally. Objective 3. Create a next-generation Beescape (Beescape NexGen) that optimizes the ability of diverse stakeholders to understand, explore, use, and contribute to information on spatiotemporal variation in pollinator health and ecosystem services. During the reporting period, the project team has designed and carried out a stakeholder evaluation to elicit ideas for Beescape NexGen features from beekeepers, growers, conservationists, and pollinator scientists. A one-day workshop held in November 2021 collected focus group data and paper prototypes for potential Beescape NexGen interfaces with a group of 21 stakeholders. A detailed report synthesizing the results of these knowledge elicitation activities will be developed in early 2022. Following the stakeholder workshop, the project team developed a new graphical mockup including interaction details for Beescape NexGen. This interface concept reflects the accumulated knowledge from usability studies of the previous Beescape prototype as well as results from our 2021 stakeholder workshop, and it will now guide the next steps of software development for Beescape NexGen. Beescape NexGen will allow users to select areas of interest using radial and polygon shapes to retrieve key pollinator habitat and health indicators. Users will also be able to change time periods to explore trends, and they will have the ability to view recent crowdsourced bee and plant observations from iNaturalist. Local climate patterns and year-to-date climate observations will also be available and associated with pollinator habitat and health indicators. Objective 4. Foster a collaborative scientific community to catalyze research in the field of bee health and landscape ecology. We are developing open-source code and repositories to allow others to readily access our data and methodologies to assess the role of land use, weather and climate impacts on bee health. The methods described in Objective 1 have been used as the basis of training materials on geospatial workflows for high performance computing. Data and methodologies are being published as stand-alone products where appropriate.
Publications
|
|