{"paper_id":"1f7ca47e-d125-497c-82e6-2a60f39c0826","body_text":"1 \nInsights from U.S. beekeeper triage surveys following unusually high honey bee colony losses \n2024-2025 \n \nAuthor list and affiliations: \nAnthony Nearman, USDA-ARS Bee Research Lab, BARC-East Bldg. 306, Beltsville, MD, 20705, USA, \nanthony.j.nearman@gmail.com, ORCID: 0000-0002-6498-3275 \nChristopher L. Crawford, USDA Office of the Chief Scientist, 1400 Independence Avenue, SW, \nWashington, D.C. 20250, christopher.crawford@usda.gov, ORCID: 0000-0001-9101-5360   \nM. Marta Guarna, Project Apis m., Salt Lake City, UT 84126, USA, marta@projectapism.org, ORCID: \n0000-0002-4638-9475 \nPriyadarshini Chakrabarti, Department of Entomology, Washington State University, Pullman, WA \n99164, USA, priya.basu@wsu.edu, ORCID: 0000-0003-0489-8586 \nKatie Lee, University of Minnesota Extension, Saint Paul, Minnesota 55108, USA, katielee@umn.edu  \nSteven Cook, USDA-ARS Bee Research Lab, BARC-East Bldg. 306, Beltsville, MD, 20705, USA, \nsteven.cook@usda.gov \nElizabeth Hill, USDA Office of the Chief Scientist, 1400 Independence Avenue, SW, Washington, D.C. \n20250, izzy@izzyhill.com   \nArathi Seshadri, USDA-ARS, Pollinator Health in Southern Crops Ecosystems Research Unit, Stoneville, \nMS 38776, arathi.seshadri@usda.gov  \nGarett Slater, Texas A&M AgriLife Extension, Overton, TX 75684, USDA garett.slater@ag.tamu.edu  \nZac Lamas, USDA-ARS Bee Research Lab, BARC-East Bldg. 476, Beltsville, MD, 20705, USA, \nzac.lamas@gmail.com  \nYan Ping Chen, USDA-ARS Bee Research Lab, BARC-East Bldg. 476, Beltsville, MD, 20705, USA, \njudy.chen@usda.gov  \n*Danielle Downey, Project Apis m., Salt Lake City, UT 84126, USA, danielle@projectapism.org \n*Jay D. Evans USDA-ARS Bee Research Lab, BARC-East Bldg. 476, Beltsville, MD, 20705, USA, \njay.evans@usda.gov \n*co-corresponding author\n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n2 \nAbstract \nIn January of 2025, U.S. commercial beekeepers reported unusually high honey bee colony \nlosses as they prepared colonies for almond pollination. Two industry groups launched \nnationwide surveys to document colony losses between June 2024 and March 2025 across all \nscales of beekeeping. This study analyzes these survey data to assess colony losses, estimate \nfinancial impacts, and identify correlations with beekeeper management practices and \ngeographical locations. Unlike past surveys, commercial beekeepers experienced more severe \nlosses than smaller-scale beekeepers during this period. Respondents, managing over half of \nU.S. colonies, most frequently cited Varroa mites as the cause for their losses. Varroa mites \nwere followed by pesticides and pathogens in the case of commercial beekeepers and by queen \nfailure and weather in the case of smaller-scale beekeepers. Although Varroa was the most \nfrequently cited cause, losses did not significantly differ between users and non-users of \namitraz, suggesting that rising amitraz resistance alone does not explain observed trends. \nDifferences in protein and carbohydrate feeding frequencies also played a role in net losses. \nWhile colony loss rates and financial concern varied widely among respondents, commercial \nbeekeepers understandably showed higher sensitivity to financial impacts, with concerns \nincreasing linearly with loss severity. This study highlights the value of beekeeper surveys \nwhich, alongside direct analyses of bee samples and longitudinal studies, help identify effective \nmanagement strategies and environmental risks. Such insights are crucial for addressing the \nleading causes of colony losses on a national scale, and ultimately aid in safeguarding honey \nbee health, pollination services, and agricultural production. \n \nKey words \nPollination, parasites, pesticides, climate, agriculture \n \nHighlights \n• Unprecedented honey bee colony losses \n• Indications of disease stress \n• High economic pain for commercial beekeepers and growers \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n3 \n1. Introduction \n Honey bees are vital players in agriculture and are important community members as they are \nboth abundant in numbers and generalist pollinators. Across the globe, honey bees play an essential \neconomic role for a wide variety of agriculturalists, ranging from small subsistence farmers to massive \ncommercial operations with tens of thousands of colonies that add hundreds of billions of U.S. dollars \nto the worldwide economy (Khalifa et al. 2021, Stahlmann-Brown et al. 2023). In the United States, \nhoney bees pollinate over 100 fruits, nuts, and vegetables, along with over 40 seed crops, sustaining \nbillion-dollar crops in almonds and other high-value crops (Jordan et al. 2021). The role of honey bees \nas crop pollinators is at risk due to multiple biotic and abiotic stressors (Traynor 2015, Steinhauer et al. \n2021, French et al. 2024), While the net number of honey bee colonies in the U.S. has remained \nconstant for decades, high colony loss rates lead to costly efforts to protect and replace colonies. \nCommunity-based surveys provide insights into annual changes in colony loss rates and can \nindicate which management practices and environmental stresses are correlated with differing colony \nsurvival success. The U.S. Beekeeping Survey has provided insights into the beekeeping industry and \ncolony loss rates since its first iteration following heavy losses in 2006-2007. Funded by the USDA \nAnimal and Plant Health Inspection Service and run first by the Bee Informed Partnership, and more \nrecently by Auburn University and the Apiary Inspectors of America (Bruckner et al. 2023), these \nsurveys show that beekeepers’ self-reported annual colony loss rates averaged approximately 40% \nover the past decade (Aurell et al. 2024). While these loss rates reflect averages, it has been apparent \nsince the start of this survey that beekeeping losses are highly variable location-to-location, year-to-\nyear, and beekeeper-to-beekeeper. In January 2025, U.S. commercial beekeepers notified the USDA-\nARS Bee Research Laboratory of unusually high honey bee colony losses as they were preparing \ncolonies to be moved for almond pollination. To investigate the scope and breadth of these claims, two \nindustry organizations (Project Apis m. and the American Beekeeping Federation) developed and \ndeployed community surveys to document colony losses, determine financial impacts, and correlate \nlosses with management practices or geographical locations. This study reflects an analysis of the \nresults of these two colony loss surveys. \nBy targeting beekeepers at all scales, these surveys generated data for over half of the \nmanaged bee colonies in the U.S. (Table 1). They recorded slightly larger losses than prior years over a \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n4 \nnine-month window, along with tremendous variance across beekeeping operations in colony losses. \nUnlike prior years, commercial beekeepers (i.e., with >500 hives) faced especially high colony losses \nbetween June of 2024 and February of 2025, compared to sideliners (i.e., beekeepers managing \nbetween 50 and 500 hives) and hobbyist beekeepers (i.e., with <50 hives). We mined these losses for \npotential causes, including mite parasitism, pathogens, nutritional stress and pesticide exposure. Since \ncommercial beekeepers have begun to adopt indoor ‘shed’ storage options as a way of increasing \noverwintering survival (Hopkins et al. 2021, Degrandi-Hoffman et al. 2023, Hopkins et al. 2023), we also \nexplored outcomes for beekeepers using this method as opposed to outdoor storage. \n Varroa mites and their associated viruses are often implicated in colony loss events (Dainat et \nal. 2012, Steinhauer et al. 2018, Stahlmann-Brown et al. 2022, Lamas et al. 2025). To investigate the \nroles of various Varroa mitigation strategies in recent losses, we correlated beekeeper reports on mite \ntreatments against colony loss rates observed by the beekeeper. We focused especially on the use of \nthe miticide amitraz, since Varroa mites have recently acquired widespread resistance in the U.S. to \nthis common treatment (Hernández-Rodríguez et al. 2022, Rinkevich et al. 2023). \nAdditionally, we examined whether the type or frequency of carbohydrate and protein \nsupplementation was correlated with colony losses. Supplemental feeding is a critical part of \nbeekeeping management practices; this is especially true for beekeepers managing colonies during \ntimes of forage scarcity, which can be especially pronounced in some regions of the U.S. (DeGrandi-\nHoffman et al. 2016, Chakrabarti et al. 2020, Tsuruda et al. 2021, Bernklau and Arathi 2023, \nChakrabarti and Sagili 2023). \nLoss rates were highly variable across operations, a result echoed by beekeepers’ \nfinancial sentiments in the past year. This mimics the pattern of loss events in 2023, wherein \nsome beekeepers lost up to $1 million U.S. dollars due to economic hardships such as \ndecreased revenue from unfulfilled pollination contracts and the costs of rebuilding colony \nnumbers (Lamas et al. 2024). Commercial beekeepers were far more sensitive to the financial \nperil of their losses, showing a linear increase in financial concern. Sideliner beekeepers showed \nheightened concern when their losses topped 40-60% while hobbyists expressed significantly \nhigher concern only when losses exceeded 80%. This survey demonstrates geographically \nwidespread high colony losses that were similar across a range of operation sizes and \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n5 \nmanagement routines. As in prior reports, losses appear to reflect a range of factors ranging \nfrom parasitic mites and other biotic factors to environmental stresses. \n \n2. Methods \n2.1 Survey design and deployment \nTwo surveys were developed in parallel in early February 2025, when it was apparent that \nseveral commercial beekeepers in the U.S. were suffering extreme colony losses. One survey \nwas driven by the industry nonprofit group Project Apis m. (PAm; \nhttps://www.projectapism.org/) and one by the industry trade group the American Beekeeping \nFederation (ABF; https://abfnet.org/). The PAm survey included 16 questions (Supplemental file \nS1) covering beekeeper operation size, geographical location(s), types of supplemental feeding, \nestimated ranges of fall mite levels, methods used to treat Varroa mites, queen replacement \nrates, methods of winter storage, perceived causes of colony loss, financial concern, and \nrealized colony losses from June 2024 to February 2025. This survey was widely advertised in \nindustry publications, online beekeeping groups, social media, and listservs dedicated to \npollination and honey bees, resulting in ‘snowball sampling’ whereby early participants also \nrecruited others. The survey ran from February 1, 2025, to March 15, 2025, capturing 842 \nunique responses (Table 1). PAm also distributed a second, follow-up survey to respondents \nwho agreed to be contacted with additional questions. The purpose of the follow-up survey was \nto gather information on individual operation sizes (i.e., number of hives), which were in turn \nused to estimate the total number of colonies lost. This follow-up survey captured 110 unique \nresponses, enabling the estimation of mean operation sizes for sideliners (mean = 314 hives, N \n= 6) and commercial beekeepers (mean = 6,798 hives, N = 90). When individual operation sizes \nwere unavailable, these mean operation sizes were combined with individually reported loss \nestimates from the primary survey to estimate the number of colonies lost (Table 1; a value of \nfive hives was assigned to hobbyists based on expert judgement, to provide a conservative \nestimate). The ABF survey presented similar questions regarding operation size, location, \nmiticide treatments, queen replacement rates, winter storage practices, and realized colony \nlosses from June 2024 to early 2025. This survey was sent to members of the American \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n6 \nBeekeeping Federation, an industry group of 1000 members, and captured 107 unique \nresponses (Table 1). \n \n2.2 Data compilation and cleaning \nComma-delimited files with survey results were collected at the end of each survey period, \nstripped of personally identifying information and curated by scientists specializing in the \nvarious query topics. In both surveys, declared summer locations of the colonies were binned at \nthe state level and subsequently grouped into U.S. Climate Zones, as delineated by NOAA (Karl \nand Koss 1984). When respondents managed colonies across multiple states, responses were \nattributed to the first state mentioned. Since several of the queries were open-ended, results \nwere cleaned and binned as described in the Supplemental Methods. Beekeepers were \nclassified according to operation size (i.e., number of hives) as follows: commercial (> 500 \nhives), sideliners (between 50-500 hives), and hobbyists (<50 hives). Cleaned datasets are made \navailable as Supplementary Material, along with metadata describing each data field and the \nmethods used to derive structured fields from the original survey questions. Metadata and \ndetails describing the data and associated cleaning and processing steps can be found in the \nSupplementary Information and Tables S4-S6. \n \n2.3 Statistical analysis \nAll analysis was carried out using R programming language v4.5.0 “How About a Twenty-Six” \n(Team 2025). For each survey question, responses were categorized as binary or binned into \nmultivariate categories either by biological relevance or statistically relevant sample sizes. Each \nquestion was then weighed against a set of common climate covariates. To achieve this, climate \ndata was collected from PRISM (Daly et al. 1997) that represented the mean daily precipitation, \ntotal precipitation, and mean temperature from June 2024 through September 2024, as well as \nthe mean minimum temperature from November 2024 through January 2025, for each U.S. \nstate in the lower 48 and merged with the larger data set. This set of covariates was used to \nweight observations separately for each class of beekeeper (Commercial, Sideliner, and \nHobbyist) using WeightIt v1.4.0 (Greifer 2019), with either the “ebal” (Hainmueller 2012, Zhao \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n7 \nand Percival 2017) or “energy” (Huling and Mak 2024) method where required to achieve a \nbalance threshold below 0.05 for all covariates. Rarely, classes of beekeepers who could not be \nbalanced within a survey question, typically due to low response rate, were dropped from \nfurther significance testing, as indicated in the figures. As raw colony count data was available \nfor only a subset of respondents, the reported proportion of colonies lost was incorporated as \nthe response variable, along with the calculated covariate-balanced weights, into each \nindividually modeled independent variable. Models were applied to a beta regression using \nbetareg v3.2-2 (Cribari-Neto and Zeileis 2010). For multivariate questions, multiple comparisons \nwere performed with emmeans v1.11.1 (Lenth et al. 2018) and compact letter display was \nmapped with multcomp v1.4-28 (Hothorn et al. 2012). All plots were generated using ggplot2 \nv3.5.2 (Kassambara et al.). \n \n3. Results \nThe PAm survey included responses from 280 commercial beekeepers (> 500 hives) and a total \nof 842 responses. Participation was balanced between small-scale beekeepers (“hobbyists” \nwith <50 hives, N = 393), “sideliners” (50-500 hives, N = 161), and larger commercial \nbeekeepers (Table 1). The ABF survey tallied responses from 87 commercial beekeepers (> 500 \nhives) and 20 smaller-scale beekeepers (< 500 hives, i.e., hobbyists and sideliners).  \n \n \n \n \n \n \n \n \n \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n8 \nSurvey Beekeeper Class \nMean Net \nLoss Rate \nNumber of \nresponses \nEstimated Number \nof Colonies Lost* \nPAm \nHobbyist (1-49 hives) 51.2% 393  974 \nSideliner (50-500 hives) 53.9% 161 26,645 \nCommercial (>500 hives) 62.3% 280 1,147,613 \nNA 40.0% 8 - \nTotal 55.3% 842 1,175,232 \nABF \nHobbyists and Sideliners (<500 \nhives) \n32.7% 20  \nCommercial (>500 hives) 41.9% 87   \nTotal 40.2% 107   \n \nTable 1. Mean net colony loss rates captured in the two industry beekeeper surveys, conducted \nby Project Apis m. (PAm) and the American Beekeeping Federation (ABF). Results are shown by \nbeekeeper class (i.e., operation size, in terms of number of hives), and across all respondents \n(“Total”). Mean values were calculated per respondent and represent the net loss rate \nexperienced by the mean beekeeper. *Estimated number of colonies lost was calculated using \ndata from PAm’s follow-up survey (see Methods).  \n \nTo ensure the quality of our covariates, we modeled each as independent variables against \neither Net Loss, Summer Loss, or Winter Loss as dependent variables. The only exception being \nmean minimum winter temperature, which was only modeled against Winter Loss. Significant \nresults are presented in Table S1 and indicate our selected weather covariates are significant \nfor predictors for Hobbyist and Sideliner beekeepers but not Commercial beekeepers. The only \nexception to this being mean daily precipitation during the active season, which was not \nsignificant for all classes of beekeepers for any period of losses. Significant associations with \ncovariates were consistent between Sideliners and Hobbyists. Increasing mean minimum winter \ntemperature was associated with lower winter losses (Figure S1). Similarly, increasing total \nactive season precipitation and increasing mean daily active season temperature were \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n9 \nassociated with decreased net loss (Figure S2-S3). We also investigated the relationship \nbetween summer and winter loss reported by beekeepers and found that summer loss is a \nsignificant predictor for winter loss among all classes of beekeepers (Figure S4).  \n \nThe PAm survey showed widespread losses across the country (Figure 1), while the ABF survey \nshowed similarly widespread losses but at generally lower loss rates. \n \nFigure 1. Mean net colony loss rates reported in two industry beekeeper surveys, shown by \noperation size (number of hives) and aggregated by region (U.S. Climate Zones, as delineated by \nNOAA; Karl and Koss 1984, link). Mean net loss rates are shown for beekeepers according to \ntheir first listed summer location and operation size according to the number of hives (noted in \nparentheses). Numbers of beekeepers in each group are shown in white boxes superimposed \non each region. \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n10 \n \nFigure 2. a) Colony loss rates by beekeeper class, PAm survey. b) Colony loss rates by miticide \ntype, PAm survey. \n \nBased on the PAm survey, commercial beekeepers suffered significantly higher losses \nthan did hobbyist beekeepers (Figure 2a, df=2, 𝜒2=11.925, Pr(>𝜒2)=0.003). \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n11 \nA total of 821 beekeepers provided information for the PAm question asking about \nproduct use and treatment frequency to control Varroa mites. Commercial beekeepers who did \nnot identify the type of chemistry used as a miticide had significantly higher net colony losses \nthan beekeepers who reported using only amitraz-based products, an amitraz-based product \nplus another chemistry, or used only non-amitraz based chemistries (Figure 2b, df=3, 𝜒2=12.78, \nPr(>𝜒2)=0.005). Beekeepers who reported using only non-amitraz products had similar losses to \nthose who reported using amitraz (Figure 2b). Hobby beekeepers who did not use a chemical \ntreatment had higher net colony losses compared to beekeepers who used non-amitraz based \nproducts (Figure 2b, df=4, 𝜒2=13.336, Pr(>𝜒2)=0.010).  \nThe most frequently used miticides among commercial beekeepers were amitraz and \noxalic acid, reported by 133 and 115 of the 275 respondents to the PAm survey, respectively. \nFor sideliners and hobbyists, oxalic acid was the most used treatment, reported by 85 of 160 \nsideliner respondents and 136 of 383 hobbyist respondents. The other commonly used \nmiticides across all beekeeper groups were formic acid and thymol (Table S2). Beekeepers \nreported a treatment frequency range of 0 to 22 times in the June to December survey period. \nCommercial beekeepers reported treating their colonies an average of 4.7 times, compared to \n3.5 times for sideliners and 2.1 times for hobbyists (Table S3). Of those who reported treating \n20 or more times, one listed oxalic acid vapor, one rotated formic or thymol every 10 days, and \nthe other three did not list the treatment used.  \nSimilar results regarding amitraz use and colony losses were obtained from the ABF \nsurvey, even though the number of respondents was lower. Respondents to the ABF survey \nwho used amitraz reported average losses comparable to those who did not use amitraz (Figure \nS5a). In addition, no differences in colony losses were observed between those who felt their \nmite treatments were effective and those who did not (Figure S5b). The most frequently used \nmiticides were amitraz alone, and amitraz & oxalic acid (Figure S6), consistent with responses \nfrom the PAm survey. Beekeepers responding to the PAm survey were also asked to provide an \nestimated range of mite levels in their colonies. The most frequently reported range across all \nthree beekeeper classes was 2-5 mites per 100 bees, although some did not test for Varroa \nlevels. Analysis of these data revealed no clear differences in colony losses between beekeepers \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n12 \nwho provided different responses. One notable exception was among sideline beekeepers: \nthose who reported not testing for Varroa experienced higher colony losses than those who \nmonitored mite levels (Figure S7). \nSupplemental feeding practices varied across regions, seasons, and beekeeper \noperation types (commercial, sideliner or hobbyist) as reported in the PAm beekeeper survey \n(Figure S8). For supplemental protein feeding, 231 commercial beekeepers responded “Yes” \nwhile 46 commercial beekeepers responded “No” to the question. For sideliner beekeepers, 95 \nresponded “Yes” and 46 responded “No” to supplemental protein feeding. In the hobbyist \nbeekeeper group, 140 respondents fed supplemental proteins to their colonies while 170 did \nnot. There was no significant difference in net colony loss reported for commercial (Figure 3a; \ndf=1, 𝜒2=3.536, Pr(>𝜒2)=0.060), sideliner (Figure 3a; df=1, 𝜒2=1.992, Pr(>𝜒2)=0.158) and \nhobbyist (Figure 3a; df=1, 𝜒2=0.949, Pr(>𝜒2)=0.330) beekeeper groups with and without \nsupplemental protein feeding. The commercial (Figure 3b; df=3, 𝜒2=0.823, Pr(>𝜒2)=0.844) and \nhobbyist (Figure 3b; df=3, 𝜒2=4.462, Pr(>𝜒2)=0.216) beekeeper operation types did not \nsignificantly differ in their net colony losses when they fed their colonies protein supplements \nat different frequencies. However, the sideliner beekeepers reported a significantly higher net \ncolony loss when they fed their colonies protein supplements only once (Figure 3b; df=3, \n𝜒2=11.960, Pr(>𝜒2)=0.008). Among beekeeper operations, 51 commercial beekeepers, 24 \nsideliner beekeepers and 47 hobbyist beekeepers fed their colonies proteins once during the \nyear. Alternatively, 63 commercial beekeepers, 30 sideliner beekeepers and 33 hobbyist \nbeekeepers fed their colonies proteins twice during the year, whereas 70 commercial \nbeekeepers, 25 sideliner beekeepers and 34 hobbyist beekeepers supplemented their colonies \nwith proteins more than three times during the year. \nFor the supplemental carbohydrate feeding survey question, 255 commercial \nbeekeepers responded “Yes” while 6 commercial beekeepers responded “No” to providing their \ncolonies with supplemental carbohydrates. For sideliner beekeepers, 127 responded “Yes” and \n16 responded “No” to supplemental carbohydrate feeding. In the hobbyist beekeeper group, \n258 respondents fed supplemental carbohydrates to their colonies while 55 did not. There was \nno significant difference in net colony loss reported for the sideliner (Figure 3c; df=1, 𝜒2=0.258, \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n13 \nPr(>𝜒2)=0.611) and hobbyist (Figure 3c; df=1, 𝜒2=0.001, Pr(>𝜒2)=0.982) beekeeper groups \ndelineated by whether they did or did not feed sugar. As only six commercial beekeepers \nresponded “no” to supplemental carbohydrate feeding, the weighted analysis was not balanced \ndue to the low sample size (Figure 3c). Altogether, 122 commercial beekeepers, 81 sideliner \nbeekeepers and 170 hobbyist beekeepers fed their colonies carbohydrates less than four times \nwhereas 111 commercial beekeepers, 37 sideliner beekeepers and 67 hobbyist beekeepers \nsupplemented their colonies with carbohydrates more than four times during this time period. \nCarbohydrate feeding frequency did not significantly affect loss rates for commercial (Figure 3d; \ndf=1, 𝜒2=1.045, Pr(>𝜒2)=0.307) or hobbyist (Figure 3d; df=1, 𝜒2=0.697, Pr(>𝜒2)=0.404) \noperations. However, sideliner beekeepers experienced a significantly lower net colony loss \nwhen they fed their colonies carbohydrate supplements more than four times between June \n2024 - March 2025 (Figure 3d; df=1, 𝜒2=4.808, Pr(>𝜒2)=0.028).  \n     \n \nFigure 3. Weighted colony loss rates by supplemental carbohydrate or protein management as \nreported in the PAm beekeeper survey.  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n14 \nOther questions addressed management practices, including the number of queen \nreplacements and whether colonies were overwintered indoors or outdoors. Analysis of these \nresponses revealed that commercial operations reporting the highest percentages of queen \nreplacement in the PAm survey also experienced higher colony losses (Figure S9). In contrast, \nsimilar losses were reported by beekeepers who used sheds for wintering their colonies and \nthose who wintered outdoors. This pattern was observed among commercial beekeepers in the \nPAm survey, and among beekeepers with more than 50 colonies in the ABF survey (Figure S10). \n \n \n \nFigure 4. Suspected causes of losses reported by beekeepers in the PAm survey, shown by \nbeekeeper class. Responding beekeepers could identify more than one possible cause of losses.  \n \nWhen prompted to report perceived causes of loss in the PAm survey, both commercial \n(n = 280) and sideliner beekeepers (n = 160) who chose a cause for their losses most frequently \nchose “Varroa” (Figure 4). For hobby beekeepers who chose a cause (n = 347), “Varroa” and \n“Weather” were chosen most frequently. Pesticides ranked second for commercial beekeepers, \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n15 \nfollowed by pathogens (brood diseases and viruses that could be Varroa associated; of the 103 \nbeekeepers who indicated pathogens, n = 27 indicated “viruses,” n = 78 indicated “disease,” \nand n = 2 chose both viruses and disease). Nine beekeepers wrote that losses were due to a \nhurricane. Beekeepers in the PAm survey could write in multiple stressors and frequently did. \nNotably, commercial beekeepers identified multiple perceived causes of colony losses more \nfrequently than other beekeeper groups (18.9% compared to 12.5% for sideliners and 8.9% for \nhobby beekeepers).  \n \n \n \n \nFigure 5. Financial concern of beekeepers compared to actualized losses, reported in the PAm \nsurvey. \nFinancial concern for commercial beekeepers trended highest across the three \noperation sizes, with concern levels rising in tandem with increasing colony losses (Figure 5). \nThirty-two percent of the commercial beekeepers indicated that on a 10-point Likert scale, their \nfinancial concerns were a ten. One of the two commercial beekeepers who ranked financial \nconcern as a 0, also put in \"No, I quit\" for the question about contacting them. A relationship \nbetween net loss and financial concern was also observed for hobbyists and sideliners, with \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n16 \nsideliners tending to show increased concern as losses went above ca. 30% while hobbyists \ntended to express heightened concern only when losses exceeded 75%. The relationship \nbetween financial concern and loss was stronger for sideliners than for hobbyists. This makes \nsense, as sideliners rely on their colonies for income more than hobbyists do, though not as \nheavily as commercial beekeepers.  \n \n4. Discussion \n Honey bees are critical to our agroecosystem but maintaining colonies year after year \nhas become an increasing challenge. Beekeepers devote a large fraction of their budgets to the \nlabor and material costs of rebuilding annual losses, reducing disease stress and providing \nsupplemental feed. Despite these investments, beekeepers can lose entire operations \nsuddenly, with devastating personal and industry-level economic impacts (Lamas et al. 2024). \nTraditional local knowledge has often been relied upon in a Multiple Evidence Based approach \nto inform decisions for management (Smith et al. 2017). Thus, beekeeper perceptions and their \ncolony management history can often help in interpreting the reasons for colony decline and \nlosses. Sudden colony loss events, as observed in 2024-2025, offer an opportunity to explore \nand strive to mitigate these potential causes. The two surveys described here, although not as \ncomplete as long-standing U.S. and worldwide colony health surveys (e.g., surveys conducted \nby the USDA National Agricultural Statistics Service), offer insights into the scope and key \nfeatures of loss events. The U.S. bee losses of 2024-2025 are striking in that they were so \nwidespread across the entire country (Figure 1). While losses were reported starting in the \nsummer of 2024, many of the most severe losses occurred as bee colonies were pulled from \nwinter storage in advance of profitable pollination events.  \n  Even though there is evidence of miticide resistance to amitraz (Hernández-Rodríguez \net al. 2022, Rinkevich et al. 2023), beekeepers who reported using other products in lieu of \namitraz had similar losses to those beekeepers who used amitraz exclusively or in combination \nwith another chemistry. In contrast, commercial beekeepers who identified a specific miticide \nproduct or products experienced significantly lower net colony losses than those who did not \nname the type of miticide used in their colonies. This suggests that colony losses may be \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n17 \ninfluenced by the efficacy of the products used and/or Varroa mite resistance to amitraz. \nAlthough the initial analysis of reported fall 2024 mite levels didn’t reveal clear associations \nwith colony losses, future investigations exploring associations between mite loads, mite \nmitigation strategies, and colony losses may be informative. Overall, these results highlight the \ncomplexities in interpreting the observed patterns and the need for field-vetted data.  \nOur choice to model weather characteristics as covariates was meant to provide higher \nresolution, in terms of local management practices and bee behaviors, compared to simple \ngeographic location. For example, stationary apiaries in colder, northern climates have shorter \nactive seasons and are exposed to fewer or different floral resources compared to their \nsouthern counterparts. Weather is also critical for honey bee success and survival (Calovi et al. \n2021, Insolia et al. 2022, Overturf et al. 2022). By removing the effects of local climate, we were \nbetter positioned to estimate the effects of reported management decisions in the survey. In \nevaluating any direct associations between our weather covariates and reported loss, we found \nsignificant associations only among sideliners and hobbyists. This may be because these groups \nare variably stationary, or completely in the case of hobbyists, and more subject to local \nweather changes. Further, commercial operations may either be migratory or span multiple \ngeographic locations, avoiding or diluting any weather-related losses. \nOur work suggests climate variables are significant among presumably stationary and \nsingle-location sideliners and hobbyists. Recent work, however, indicates that regions \ncommercial beekeepers rely upon for staving off the effects of dearth or climate may be \nexperiencing extremes regarding drought, precipitation, or changes to the landscape tending \ntowards being more crop-dominated (Morton et al. 2015, Ahlering et al. 2020, Hoell et al. 2021, \nHaigh et al. 2022, Heim Jr et al. 2023). While extreme drought and precipitation may have \nmeasurable effects on crop yields and pollinator friendly volatiles (Kuppler and Kotowska 2021) \nand products (Vaudo et al. 2024), the downstream effects on pollinator health appears to vary \nper location and pollinator, with some work indicating a range of tolerance, that may relate to \nchanges in landscape (Descamps et al. 2021, Brunet et al. 2025). Our data suggest the locations \nand subsequent losses reported by commercial beekeepers are unrelated to climate variables, \nbut should that change under future surveillance efforts, those data would be invaluable to \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n18 \nidentifying the locations and conditions that lead to increased losses for commercial \nbeekeepers. As such, loss surveys should continue to monitor any relationships to climate-\nrelated variables and colony losses moving forward. Additional work may also be performed on \nthe data published here to determine any location-specific effects among commercial \nbeekeepers, though the data does not discern between migratory and multi-state beekeepers \nand their respective losses in each location. \nWe also found that reported summer losses predicted winter losses, an effect that was \nmost prominent among commercial beekeepers (Figure S4). This appears to be driven by \nbeekeepers who experience all around lower losses (<25%), rather than a one-to-one \ncorrelation. Still, the data indicate that there are beekeepers who experience variable rates of \nsummer and winter loss, outside of only low or only high net losses. Additional analyses should \ninvestigate any relationships between management decisions and the variability of loss across \nall seasons.  \nSupplemental feeding is often an important part of beekeeping management and the \nbeekeepers’ location, weather conditions, participating in pollination services and colony needs \nmay drive the demands for supplemental protein and carbohydrate feeding throughout the \nyear. Sideliner beekeepers reported a significantly higher net colony loss when they \nsupplemented their colonies with proteins only once and carbohydrates less than four times \nannually, compared to other groups. This reiterates the need for further assessments of local \nforage availability, colony densities and weather conditions for colonies which may drive their \ndependence on supplemental feeding. \nBeekeepers self-reporting causes associated with losses most frequently chose Varroa, \nconsistent with past beekeeper loss surveys (Aurell et al. 2024). However, other contributors \nwere also identified as being important, including pesticides, viruses, nutrition, or regional \nweather events. Varroa mite levels are immediately measurable, while contributions by other \nfactors, like pesticides and nutrition, can be less apparent. In addition, colonies weakened by \nfactors such as pesticides and poor nutrition may become susceptible to Varroa mite \ninfestation, obscuring the primary cause. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n19 \n While other surveys have measured high honey bee losses, including surveys conducted \nby the U.S. Department of Agriculture's National Agricultural Statistics Service (NASS; \nhttps://www.nass.usda.gov/Surveys/Guide_to_NASS_Surveys/Bee_and_Honey/) and Auburn \nUniversity, the described PAm and ABF surveys were critical in rapidly quantifying the scope of \nspecific in-field issues experienced by beekeepers. Responses to the PAm survey included 280 \ncommercial beekeepers, while the commercial responses to three recent years of the U.S. \nBeekeeping Survey averaged only 41 commercial beekeeper respondents (Bruckner et al. 2023). \nSelf-reporting, optional surveys can be prone to response bias, and it is possible that the \ncommercial beekeepers who responded to the PAm and ABF surveys were more likely to have \nexperienced the higher loss rates. However, with over 50% of U.S. colonies represented by \nparticipants, higher than all prior surveys, the results are likely to strongly reflect the industry. \nThese surveys demonstrate the financial strain faced by commercial beekeepers, as evidenced \nby 32% of all commercial beekeepers responding to the PAm survey ranking their financial \nconcern at the highest level. These sociological indicators, coupled with in-field measurements \nof loss events (Lamas et al. 2024), demonstrate the precarious nature of bee pollination, a key \nagricultural service. \n \n5. Acknowledgements \nWe are deeply grateful to the beekeepers who took the time to respond to these surveys. CLC \nwas supported by an AAAS Science & Technology Policy Fellowship served at the USDA Office of \nthe Chief Scientist. \n \n6. Contributions \nConceptualization: Jay Evans, Danielle Downey \nData curation: Priyadarshini Chakrabarti, Steven Cook, Christopher L. Crawford, Katie Lee, \nAnthony Nearman \nFormal analysis: Anthony Nearman, Christopher L. Crawford \nFunding acquisition: N/A \nInvestigation: Jay Evans, M. Marta Guarna, Danielle Downey \nMethodology: Anthony Nearman, Christopher L. Crawford, Priyadarshini Chakrabarti, Katie Lee \nVisualization: Christopher L. Crawford, Anthony Nearman \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n20 \nWriting – original draft: Jay Evans, Priyadarshini Chakrabarti, M. Marta Guarna, Katie Lee, \nAnthony Nearman \n \n \n7. References \n \nAhlering, M. A., M. Cornett, K. Blann, M. White, C. Lenhart, C. Dixon, M. R. Dudash, L. Johnson, \nB. Keeler, and B. Palik. 2020. A conservation science agenda for a changing Upper \nMidwest and Great Plains, United States. Conservation Science and Practice 2:e236. \nAurell, D., S. Bruckner, M. Wilson, N. Steinhauer, and G. R. Williams. 2024. A national survey of \nmanaged honey bee colony losses in the USA: Results from the Bee Informed \nPartnership for 2020–21 and 2021–22. Journal of Apicultural Research 63:1-14. \nBernklau, E., and H. S. Arathi. 2023. Seasonal patterns of beneficial phytochemical availability in \nhoney and stored pollen from honey bee colonies in large apiaries. Journal of Economic \nEntomology 116:1069-1077. \nBruckner, S., M. Wilson, D. Aurell, K. Rennich, D. vanEngelsdorp, N. Steinhauer, and G. R. \nWilliams. 2023. A national survey of managed honey bee colony losses in the USA: \nresults from the Bee Informed Partnership for 2017–18, 2018–19, and 2019–20. Journal \nof Apicultural Research 62:429-443. \nBrunet, J., D. W. Inouye, E. E. Wilson Rankin, and T. C. Giannini. 2025. Global change aggravates \ndrought, with consequences for plant reproduction. Annals of Botany 135:89-104. \nCalovi, M., C. M. Grozinger, D. A. Miller, and S. C. Goslee. 2021. Summer weather conditions \ninfluence winter survival of honey bees (Apis mellifera) in the northeastern United \nStates. Scientific reports 11:1553. \nChakrabarti, P., H. M. Lucas, and R. R. Sagili. 2020. Novel Insights into Dietary Phytosterol \nUtilization and Its Fate in Honey Bees (Apis mellifera L.). Molecules 25. \nChakrabarti, P., and R. R. Sagili. 2023. Managed foraging for honey and crop pollination—Honey \nbees as livestock. Pages 175-193  The Foraging Behavior of the Honey Bee (Apis \nmellifera, L.). \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n21 \nCribari-Neto, F., and A. Zeileis. 2010. Beta regression in R. Journal of statistical software 34:1-\n24. \nDainat, B., J. D. Evans, Y. P. Chen, L. Gauthier, and P. Neumann. 2012. Dead or alive: Deformed \nwing virus and Varroa destructor reduce the life span of winter honeybees. Applied and \nEnvironmental Microbiology 78:981-987. \nDaly, C., G. Taylor, and W. Gibson. 1997. The PRISM approach to mapping precipitation and \ntemperature. Pages 20-23 in Proc., 10th AMS Conf. on Applied Climatology. \nDeGrandi-Hoffman, G., Y. Chen, R. Rivera, M. Carroll, M. Chambers, G. Hidalgo, and E. W. de \nJong. 2016. Honey bee colonies provided with natural forage have lower pathogen loads \nand higher overwinter survival than those fed protein supplements. Apidologie 47:186-\n196. \nDegrandi-Hoffman, G., V. Corby-Harris, H. Graham, E. Watkins-Dejong, M. Chambers, and L. \nSnyder. 2023. The survival and growth of honey bee (Hymenoptera: Apidae) colonies \noverwintered in cold storage: the effects of time and colony location. Journal of \nEconomic Entomology 116:1078-1090. \nDescamps, C., M. Quinet, and A.-L. Jacquemart. 2021. The effects of drought on plant–\npollinator interactions: What to expect? Environmental and Experimental Botany \n182:104297. \nFrench, S. K., M. Pepinelli, I. M. Conflitti, A. Jamieson, H. Higo, J. Common, E. M. Walsh, M. \nBixby, M. M. Guarna, S. F. Pernal, S. E. Hoover, R. W. Currie, P. Giovenazzo, E. Guzman-\nNovoa, D. Borges, L. J. Foster, and A. Zayed. 2024. Honey bee stressor networks are \ncomplex and dependent on crop and region. Current Biology 34:1893-1903.e1893. \nGreifer, N. 2019. Weighting for covariate balance in observational studies. \nHaigh, T. R., J. A. Otkin, M. Woloszyn, D. Todey, and C. Felkley. 2022. Meeting the drought \ninformation needs of Midwest perennial specialty crop producers. Journal of Applied \nMeteorology and Climatology 61:839-855. \nHainmueller, J. 2012. Entropy balancing for causal effects: A multivariate reweighting method \nto produce balanced samples in observational studies. Political analysis 20:25-46. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n22 \nHeim Jr, R. R., D. Bathke, B. Bonsal, E. W. Cooper, T. Hadwen, K. Kodama, D. McEvoy, M. Muth, \nJ. W. Nielsen-Gammon, and H. R. Prendeville. 2023. A review of user perceptions of \nDrought indices and indicators used in the Diverse climates of North America. \nAtmosphere 14:1794. \nHernández-Rodríguez, C. S., S. Moreno-Martí, G. Almecija, K. Christmon, J. D. Johnson, M. \nVentelon, D. vanEngelsdorp, S. C. Cook, and J. González-Cabrera. 2022. Resistance to \namitraz in the parasitic honey bee mite Varroa destructor is associated with mutations \nin the β-adrenergic-like octopamine receptor. Journal of Pest Science 95:1179-1195. \nHoell, A., T. W. Ford, M. Woloszyn, J. A. Otkin, and J. Eischeid. 2021. Characteristics and \npredictability of midwestern United States drought. Journal of Hydrometeorology \n22:3087-3105. \nHopkins, B. K., P. Chakrabarti, H. M. Lucas, R. R. Sagili, and W. S. Sheppard. 2021. Impacts of \nDifferent Winter Storage Conditions on the Physiology of Diutinus Honey Bees \n(Hymenoptera: Apidae). Journal of Economic Entomology 114:409-414. \nHopkins, B. K., J. Long, N. L. Naeger, and W. S. Sheppard. 2023. Comparison of indoor \n(refrigerated) versus outdoor winter storage of commercial honey bee, Apis mellifera \n(Hymenoptera: Apidae), colonies in the Western United States. Journal of Economic \nEntomology 116:1063-1068. \nHothorn, T., F. Bretz, P. Westfall, and R. M. Heiberger. 2012. Multcomp: simultaneous inference \nfor general linear hypotheses. UR L http://CRAN. R-project. org/package= multcomp, R \npackage version:1-2. \nHuling, J. D., and S. Mak. 2024. Energy balancing of covariate distributions. Journal of Causal \nInference 12:20220029. \nInsolia, L., R. Molinari, S. R. Rogers, G. R. Williams, F. Chiaromonte, and M. Calovi. 2022. Honey \nbee colony loss linked to parasites, pesticides and extreme weather across the United \nStates. Scientific reports 12:20787. \nJordan, A., H. M. Patch, C. M. Grozinger, and V. Khanna. 2021. Economic Dependence and \nVulnerability of United States Agricultural Sector on Insect-Mediated Pollination Service. \nEnvironmental science & technology 55:2243-2253. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n23 \nKarl, T. R., and W. J. Koss. 1984. Regional and national monthly, seasonal, and annual \ntemperature weighted by area, 1895-1983., National Climatic Data Center, Asheville, \nNC. \nKassambara, A., M. Kosinski, P. Biecek, and S. Fabian. Kassambara, A., Kosinski, M., Biecek, P. & \nFabian, S (2021). survminer: Drawing Survival Curves using ‘ggplot2’. R package version \n0.4.9 https://CRAN.R-project.org/package=survminer. \nKhalifa, S. A. M., E. H. Elshafiey, A. A. Shetaia, A. A. A. El-Wahed, A. F. Algethami, S. G. \nMusharraf, M. F. AlAjmi, C. Zhao, S. H. D. Masry, M. M. Abdel-Daim, M. F. Halabi, G. Kai, \nY. Al Naggar, M. Bishr, M. A. M. Diab, and H. R. El-Seedi. 2021. Overview of bee \npollination and Its economic value for crop production. Insects. \nKuppler, J., and M. M. Kotowska. 2021. A meta-analysis of responses in floral traits and flower–\nvisitor interactions to water deficit. Global Change Biology 27:3095-3108. \nLamas, Z. S., Y. Chen, and J. D. Evans. 2024. Case Report: Emerging Losses of Managed Honey \nBee Colonies. Biology 13. \nLamas, Z. S., F. Rinkevich, A. Garavito, A. Shaulis, D. Boncristiani, E. Hill, Y. P. Chen, and J. D. \nEvans. 2025. Viruses and vectors tied to honey bee colony losses. \nBioRxiv:2025.2005.2028.656706. \nLenth, R., H. Singmann, J. Love, P. Buerkner, and M. Herve. 2018. Emmeans: Estimated marginal \nmeans, aka least-squares means. R package version 1:3. \nMorton, L. W., J. Hobbs, J. G. Arbuckle, and A. Loy. 2015. Upper Midwest climate variations: \nFarmer responses to excess water risks. Journal of environmental quality 44:810-822. \nOverturf, K. A., N. Steinhauer, R. Molinari, M. E. Wilson, A. C. Watt, R. M. Cross, G. R. Williams, \nand S. R. Rogers. 2022. Winter weather predicts honey bee colony loss at the national \nscale. Ecological Indicators 145:109709. \nRinkevich, F. D., S. Moreno-Martí, C. S. Hernández-Rodríguez, and J. González-Cabrera. 2023. \nConfirmation of the Y215H mutation in the β2-octopamine receptor in Varroa \ndestructor is associated with contemporary cases of amitraz resistance in the United \nStates. Pest Management Science 79:2840-2845. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint \n\n24 \nSmith, B. M., P. Chakrabarti, A. Chatterjee, S. Chatterjee, U. K. Dey, L. V. Dicks, B. Giri, S. Laha, R. \nK. Majhi, and P. Basu. 2017. Collating and validating indigenous and local knowledge to \napply multiple knowledge systems to an environmental challenge: A case-study of \npollinators in India. Biological Conservation 211:20-28. \nStahlmann-Brown, P., R. J. Hall, R. Butt, B. McCall, G. Torres, and T. Wright. 2023. Valuing over-\nwinter colony losses for New Zealand’s commercial beekeepers. New Zealand Economic \nPapers 57:184-190. \nStahlmann-Brown, P., R. J. Hall, H. Pragert, and T. Robertson. 2022. Varroa appears to drive \npersistent increases in New Zealand colony losses. Insects 13. \nSteinhauer, N., K. Kulhanek, K. Antúnez, H. Human, P. Chantawannakul, M. P. Chauzat, and D. \nvanEngelsdorp. 2018. Drivers of colony losses. Current Opinion in Insect Science 26:142-\n148. \nSteinhauer, N., D. vanEngelsdorp, and C. Saegerman. 2021. Prioritizing changes in management \npractices associated with reduced winter honey bee colony losses for US beekeepers. \nScience of the Total Environment 753. \nTeam, R. C. 2025. R: A language and environment for statistical computing. R Foundation for \nStatistical Computing. \nTraynor, K. 2015. The complicated story of colony losses and pesticides. American Bee Journal \n155:669-671. \nTsuruda, J. M., P. Chakrabarti, and R. R. Sagili. 2021. Honey Bee Nutrition. Veterinary Clinics of \nNorth America - Food Animal Practice 37:505-519. \nVaudo, A. D., L. A. Dyer, and A. S. Leonard. 2024. Pollen nutrition structures bee and plant \ncommunity interactions. Proceedings of the National Academy of Sciences \n121:e2317228120. \nZhao, Q., and D. Percival. 2017. Entropy balancing is doubly robust. Page 20160010. De Gruyter. \n \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted August 9, 2025. ; https://doi.org/10.1101/2025.08.06.668930doi: bioRxiv preprint","source_license":"Public-Domain","license_restricted":false}