Search

High-altitude and alpine areas are predicted to experience rapid and substantial increases in future temperature, which may have serious impacts on soil carbon, nutrient and soil fauna. Here we report the impact of 20 years of experimental warming on soil properties and soil mites in three contrasting plant communities in alpine/subarctic Sweden. Long-term warming decreased juvenile oribatid mite density, but had no effect on adult oribatids density, total mite density, any major mite group or the most common species. Long-term warming also caused loss of nitrogen, carbon and moisture from the mineral soil layer in mesic meadow, but not in wet meadow or heath or from the organic soil layer. There was a significant site effect on the density of one mite species, Oppiella neerlandica, and all soil parameters. A significant plot-scale impact on mites suggests that small-scale heterogeneity may be important for buffering mites from global warming. The results indicated that juvenile mites may be more vulnerable to global warming than adult stages. Importantly, the results also indicated that global warming may cause carbon and nitrogen losses in alpine and tundra mineral soils and that its effects may differ at local scale.Other Information Published in: Scientific Reports License: https://creativecommons.org/licenses/by/4.0See article on publisher's website: http://dx.doi.org/10.1038/srep44489

Global change is causing changes in temperature, precipitation, and nutrient availability. Alpine and polar regions are predicted to be among the most vulnerable for these changes. We applied a seven-year factorial experiment with warming and nutrient addition in two alpine vegetation communities. We analyzed relationship between fruit production and monthly mean, max and min temperatures during the fall of the pre-fruiting year, fruiting summer, whole fruit production period, and effects of summer and winter precipitation on fruit production. Nutrient addition and combined nutrient addition and warming increased total fruit production, and fruit production of graminoids in the later years in the Heath and Meadow and had a positive effect on forbs in the Meadow. In contrast fruit production of evergreen shrubs were negative affected by all treatments in the Meadow and Heath, while fruit production of deciduous shrubs was negatively affected in the Meadow. Minimum and mean temperatures were more important than max temperatures, and max temperatures of the fall before flowering was more important than max temperature during the flowering year. Winter precipitation had significant effect on total fruit production, fruit production of deciduous shrubs and forbs in the Heath, and on evergreen shrubs and forbs in the Meadow. Increased nutrient availability increased fruit production over time in contrasting high alpine plant communities, while experimental warming had no, or negative effect. Additionally, the results indicate that warmer summers may have limited impact on fruit production of high alpine plants. Instead, max temperatures during the fall before the fruiting year, and minimum temperatures may be more important. This study was supported by Carl Tryggers stiftelse för vetenskaplig forskning and Qatar Petroleum to JMA.

Litter decomposition is a key driver of ecosystem processes and carbon cycling. Decomposition rate is influenced by numerous factors, such as temperature, humidity, litter properties, soil properties, and properties of soil fauna/microbial communities. The aim of this review was to summarize current knowledge on litter decomposition above the treeline in alpine regions worldwide and identify: I) factors that have been studied in great detail, II) factors that have been less intensively investigated, III) geographical regions that have been less well studied, and IV) factors with consistent or inconsistent effects on decomposition. The review showed inconsistent results for all factors covered by two or more studies regarding their effect on decomposition rate (positive, negative, no effect), usually a result of interactions between factors. Studies examining one or several factors in the physical environment (i.e., altitude, experimental warming, microclimate, snow cover and soil moisture) were most common, while studies on different aspects of resource quality were the second most common. The impacts of trophic interactions on soil microbes and fauna were less frequently studied. Europe and Asia were the best-represented regions, in terms of number of studies and geographical distribution, while there were no studies from Africa and very few from South America and Australia. North American studies were all from Colorado, and those from Asia were all from China. In order to obtain better global representation, there is a need for studies in Africa, South America, and Australia. There is also a need for more studies to explain the large variation in responses of litter decomposition rates to different influencing factors in alpine environments. Future research should focus on interactions between different factors and on experiments testing specific relationships, such as the potential interaction between temperature and soil moisture and its effect on litter decomposition above the treeline in alpine regions. This study was supported by a grant from Qatar Petroleum ( QUEX-CAS-QP-RD-18/19 ).

Uniformity of lighting for pedestrians is often assumed to have been achieved in mixed traffic environments when the lighting uniformity requirements for vehicular traffic have been fulfilled. Uniformity of lighting for drivers is commonly evaluated based on quantitative data on parameters such as overall luminance uniformity. However, methods for evaluating uniformity from the perspective of other road users are currently somewhat lacking. This study discusses qualitative and quantitative methods of assessing street lighting uniformity, and the potential implications for lighting design and the road users. We used convergence design and imbedded design based on two field studies. The research purpose is twofold: first, to study if, and how, measured lighting uniformity corresponds with visual perception. Secondly, to identify and explain the additional value that a combined method approach can contribute. The study considers examples of when the measured uniformity corresponds to visually perceived uniformity and when they do not correspond. Conference held online.

Funding Information: This study was supported by a grant to SL from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie, Grant No. 797446 and by the Independent Research Fund Denmark, Grant no. 0135-00140B. Funding from the Academy of Finland (grant 322266), National Science Foundation (1504224, 1836839, PLR-1504381 and PLR-1836898), Independent Research Fund Denmark (9040-00314B), Moscow State University, (project No 121032500089-1), Natural Sciences and Engineering Research Council of Canada, ArcticNet, Polar Continental Shelf Program, Northern Science Training Program, Polar Knowledge Canada, Royal Canadian Mounted Police, Tomsk State University competitiveness improvement program and the Russian Science Foundation (grant No 20-67-46018) are gratefully acknowledged. Matthias Ahrens provided valuable insights on the cushion growth form, and we are most thankful. We thank Gaius Shaver and two anonymous reviewers for providing valuable critique and input to earlier versions of this manuscript. Publisher Copyright: © the author(s) or their institution(s). The relative contribution of bryophytes to plant diversity, primary productivity, and ecosystem functioning increases towards colder climates. Bryophytes respond to environmental changes at the species level, but because bryophyte species are relatively difficult to identify, they are often lumped into one functional group. Consequently, bryophyte function remains poorly resolved. Here, we explore how higher resolution of bryophyte functional diversity can be encouraged and implemented in tundra ecological studies. We briefly review previous bryophyte functional classifications and the roles of bryophytes in tundra ecosystems and their susceptibility to environmental change. Based on shoot morphology and colony organization, we then propose twelve easily distinguishable bryophyte functional groups. To illustrate how bryophyte functional groups can help elucidate variation in bryophyte effects and responses, we compiled existing data on water holding capacity, a key bryophyte trait. Although plant functional groups can mask potentially high interspecific and intraspecific variability, we found better separation of bryophyte functional group means compared with previous grouping systems regarding water holding capacity. This suggests that our bryophyte functional groups truly represent variation in the functional roles of bryophytes in tundra ecosystems. Lastly, we provide recommendations to improve the monitoring of bryophyte community changes in tundra study sites.

Viltolyckor i Sverige är ett växande problem som leder till dödade och skadade personer, liksom skador på fordon och dödade djur. Viltolyckor har samhällsekonomiska kostnader som behöver kvantifieras för att man ska kunna räkna ut kostnadsnyttan av åtgärder som förebygger viltolyckor, såsom exempelvis viltstängsel eller passager. Effektmodellen för viltolyckor utvecklades under 1980-talet och ingår i EVA (effekter vid väganalyser) men är idag till vissa delar föråldrad på grund av utvecklingen som skett avseende trafiksäkerhet, viltolyckssammansättning, datasystem, kunskap och forskning. Den existerande effektmodellen bygger på risk för viltolyckor per km väg per län i tre trafikflödesklasser för älg och summerat ihop för rådjur/ren. Detta projekt har haft som mål att uppdatera effektmodellen för viltolyckor genom att inkludera aktuella data på viltolycksfördelningar och mörkertal, beakta flera viltslag (älg, rådjur, hjort och vildsvin) och involvera flera variabler som kan vara av betydelse för olyckorna: t.ex. viltets populationsstorlek, vägstandard, trafikarbete, hastighet, viltstängsel, landskapstyp och viltets populationsstorlek. Arbetet har delats upp i tre huvudsakliga delar. Den första delen omfattade att ta fram ny kunskap om sambanden mellan viltolyckor, trafikarbete, och viltpopulationer där också landskapsvariabler ingår samt kostnadsberäkningar baserat på olika typer av prognoser på längre sikt. En annan del omfattade att ta fram ny viltolycksstatistik baserat på GIS-analyser för uppdaterade effektsamband för viltolyckor (olycksfrekvens per km väg), ÅDT1 , vägstandard och stängsel samt nya skattningar av mörkertal. Den tredje delen omfattade att ta fram nya värden för skadeföljd och skadekonsekvenser för olika viltslag för 2010–2016. Resultaten från första delen av projektet visar att utifrån ett oförändrat jakttryck kommer viltolyckorna att öka med 12% från 2015 fram till 2030, mestadels på grund av en fördubbling av vildsvinsolyckor medan risken för olycka med älg och rådjur minskar något. Vid minskat jakttryck under samma period visar resultaten att det är teoretiskt möjligt att viltolyckorna istället kommer att öka med 50%. Beräknade kostnader för viltolyckor år 2030 (ASEK 6) varierar mellan 1 707 och 7 556 miljoner kronor beroende på antaganden om jakttryck och skadekostnader. Olyckor med rådjur svarar för den största delen av olyckorna under hela perioden, men olyckor med vildsvin ökar mest under samtliga scenarier. Prognos för kostnadsberäkningar av viltolyckor för Västra Götaland och Skåne inkluderas som exempel. Data från 2010–2016 visar att viltolyckorna ökat mer eller mindre konstant sen 2003 och att ökningen i olyckstal över åren sker med snabbare takt än ökningen i trafikarbetet. På mindre vägar tilltar olycksfrekvenser med ökande trafikarbete fram till en genomsnittlig trafikvolym av omkring 5 000 ÅDT. På större vägar avtar däremot olycksfrekvensen med stigande trafik. Olycksfrekvensen är högst på mellanstora vägar men avtar på trafiktunga vägar och på vägar med hög hastighet. Uppdaterade värden för skadeföljd och skadekonsekvens baserades på uttag ur STRADA2 och NVR3 2010–2016 och endast polisrapporterade olyckor (förutom för RPMI4 värden där även sjukhusrapporter ingick). Skadeföljden visar att dödade och skadade personer i älg- och rådjur/hjortolyckor har minskat och justeringar i de nya värdena har gjorts. För skadekonsekvenser föreslås nya värden för älgolyckor medan föregående värden behållits för rådjur/hjort och även föreslås för vildsvin. Nya värden för skadeföljd föreslås för rådjur/hjort och vildsvin. Slutligen diskuteras antagande i effektmodellen och resultat utifrån trafiksäkerhetsutvecklingen, relevanta viltolycksfrågor, samt några utvalda aspekter inför effektmodellens implementering. Wildlife-vehicle accidents in Sweden are a growing problem that causes killed and injured humans, damage to vehicles and killed animals. Wildlife-vehicle accidents have socioeconomic costs that need to be quantified to calculate the cost-benefits of measures to prevent the accidents, for example game fencing or animal passages. The effect-model for wildlife-vehicle accidents was developed in the 1980s but is currently obsolete due to developments in road safety, changes in wildlife populations, data systems, knowledge and research. The existing effectmodel is based on the risk of accidents per km road per county in three traffic flow categories and divided into moose and summed up for deer and reindeer. This project aims at updating the effect-model for wildlife-vehicle accidents by including current data on wildlife collisions, considering various species of wildlife (elk, deer and wild boar) and several variables that may be of significance to the accidents: such as: population size, road standard, traffic load, vehicle speed, and landscape type. The work has mainly been divided into three parts. The first part involved the development of a model including wildlife accidents, traffic load and population size/growth and landscape variables, as well as cost estimates based on different model scenarios in long-term forecasts. Another part involved GIS analyzes of updated wildlife-vehicle accident statistics; accident rate per km road, AADT5 , road standard and fencing, and new estimates of unknown cases. The third part comprised presenting new values for damage and risk consequences (percentage probabilities for fatalities and personal injuries) for vehicle accidents in recent years. The results for the first part of the project show that wildlife-vehicle accidents will increase by 12% from 2015 to 2030, mainly due to a doubling of wild boar accidents, whereas the risk of accident with elk and deer decreases slightly. In the case of reduced hunting pressure, the results show that it is theoretically possible that wildlife accidents will instead increase by 50%. Estimated costs of game accidents in 2030 (ASEK 6) vary between 1 707 and 7 556 million SEK, depending on assumptions about hunting and injury costs. Accidents with deer are responsible for most of the accidents throughout the period, but accidents with wild boar increase most in all scenarios. Cost estimates of wildlife accidents for Västra Götaland and Skåne in a long-term forecast are included as examples. The GIS analyzes on data from 2010-2016 show that wildlife-vehicle accidents has increased constantly since 2003 and that the increase in accidents over the years occurs at a faster rate than the increase in traffic load. On minor roads accidents will increase synchronously with the traffic volume up to approximately 5 000 AADT. On larger roads, accident frequency will decline with increased traffic load. The accident frequency is highest on medium-sized roads but decreases on traffic-heavy roads and on high-speed roads. Updated values for damage consequences were based on data from STRADA6 and NVR7 2010-2016 and police reported accidents (except for calculating RPMI8 values). The results show that killed and injured persons in moose and roe deer/deer accidents have decreased and adjustments in the new values are suggested. For damage consequences, new values for moose accidents are suggested while previous values have been kept for deer and are proposed to use wild boar. Finally, assumptions in the effect-model and results based on road safety development, relevant wildlife issues, and some selected aspects regarding the implementation of the effect-model are discussed.

Rapporten är slutrapport för projektet ”LED-belysningens effekter på djur och natur med rekommendationer” (internt projektnummer AJD0008) som utförts under 2016–2018 och som genomförts av Calluna AB på uppdrag av Trafikverket och Transportstyrelsen.

This study analysed the preparedness for climate change adaptation (CCA) of transport infrastructure in eight municipalities in Sweden. The study focused on municipal officials’ role in organizing and implementing CCA measures. The interviews were done in 2011. Many respondents confused CCA with climate change mitigation. Three of the municipalities had performed vulnerability analyses but apart from that, well-documented decision support in long-term adaptation was largely lacking. Adaptation measures were often based on day-to-day problems and recently experienced weather incidents. Strategic work with long-term CCA seemed to be largely lacking. The placing of the responsibility for CCA in the municipal organization was often unclear. Six of the respondents found there were no actual hindrances to their work with CCA and no conflicts with other municipal aims. However, several responses indicated budgetary competition with environmental aims or other societal strivings. The planning and implementation of CCA measures seemed to be highly dependent on individual officials, their engagement, their passivity/activity regarding this issue, their defined responsibility and the boundaries for their current position in the municipality organization. I denna studie analyserades beredskapen för transportsystemets anpassning till klimatförändring i åtta svenska kommuner. Studien fokuserade på kommunala tjänstemäns roll i arbetet med att organisera och genomföra åtgärder för klimatanpassning. Intervjuerna genomfördes 2011. Många av de intervjuade blandade samman klimatanpassning med åtgärder mot klimatförändring. Tre av kommunerna hade utfört sårbarhetsanalyser, men väldokumenterade beslutsstöd för anpassningsåtgärder saknades i hög grad. Där åtgärder genomförts, var de ofta baserade på dagliga problem och erfarenheter från nyligen inträffade väderincidenter. Strategiskt arbete med klimatanpassning verkade i stort sett saknas. Ansvaret för klimatanpassning var ofta oklart. Sex av de intervjuade angav att det inte fanns några egentliga hinder för deras arbete med klimatanpassning eller konflikter med andra kommunala intressen. Flera svar angav dock budgetkonkurrens med kommunala prioriteringar inriktade mot miljö eller andra samhällsintressen. Såväl de intervjuades kunskaper som planeringen och genomförandet av klimatanpassningsåtgärder verkade vara starkt beroende av enskilda tjänstemän och deras engagemang samt definition och avgränsning av deras ansvarsområden inom organisationen. Planering och beslutsprocesser för klimatanpassning av drift och underhåll

En av de vanligaste orsakerna till vägtrafikolyckor är påkörning av vilda djur, i synnerhet klövdjur. Över 45 000 viltolyckor per år registrerades under de sen-aste 5 åren och antalen ökar stadigt. Trots upprepade informationskampanjer, omfattande investering i viltstängsel och andra förebyggande åtgärder och i kon-trast till avskjutningsstatistik för älg och rådjur, så ökar olyckstalen i snabbare takt än vad som kan förväntas på grund av ökat trafikarbete och större fordons-flotta. Det är tydligt att viltolycksproblematiken i Sverige inte är under kontroll. En bidragande faktor är att inte alla inträffade olycksfall rapporteras och inte alla rapporterade händelser återges av den officiella statistiken. Uppräknings-faktorerna som Trafikverket använder i sina Effektberäkningar vid Väganalyser (EVA-kalkyler) är utdaterade och utgår från effektsamband som troligtvis inte längre är aktuella. För att planera effektiva och riktade åtgärder behövs bättre kunskap om var, när och hur ofta viltolyckor inträffar. Föreliggande projekt avser att uppdatera kunskaperna om mörkertalet i rappor-teringen av viltolyckor på väg och belysa andra bortfall i statistiken. Målsätt-ningen är att ta fram ett uppdaterat underlag (Uppräkningsfaktorer) för använd-ning i effektberäkningar vid väganalyser samt undersöka kvarstående utred-ningsbehov och eventuella förbättringsmöjligheter som kan leda till en bättre uppskattning av konflikten mellan vilt och vägtrafik. Projektet omfattar tre delstudier: enkätundersökning med bilförare, analys av viltrelaterade personskadeolyckor i STRADA (d.v.s. Swedish Traffic Accident Data Acquisition) samt en jämförande analys av olycksrapporter från polisens och eftersöksjägarnas inrapportering i viltolycksstatistiken tillgänglig hos NVR (Nationella Viltolycksrådet). En enkätstudie via internet (på domänen www.viltochtrafik.se) genomfördes för att undersöka storleken på mörkertalet i rapportering av viltolyckor från bilfö-rare till polis. Enkäten var tillgänglig för allmänheten online under perioden 2013-10-31 till 2014-12-31 och besvarades av 3981 respondenter. Vi ställde frå-gor kring kunskaper om och erfarenheter med viltolyckor och bad om mer detal-jerade uppgifter om den incidenten som inträffade för respondenterna efter 2004. Mer än hälften av alla respondenter (65 %) har varit med om en klövvilt-olycka någon gång och runt 20 % av dessa angav att olyckan troligen aldrig rap-porterats till polisens kännedom. Drygt 45 % av respondenterna angav att de va-rit med om viltolyckor efter år 2004 och mörkertalet uppskattades till mellan 9 % och 19 %. Vi rekommenderar att det schablonmässiga mörkertalet för klövvilt (oavsett art) bör vara 15 % (±5 %). Våra resultat visar att mörkertalet i rapporte-ringen från bilist till polisen har närmare halverats sedan 1970-talet. Efter att en olycka anmälts till polisen finns risk för ytterligare bortfall i statisti-ken. Inkommande polisanmälningar överförs manuellt till olika databaser, bero-ende på hur händelsen klassificeras. Olyckor där viltet inte utgjorde den omedel-bara orsaken till skadan klassas inte som viltolycka och återfinns då inte i den officiella polisiära statistiken på viltolyckor. Detta bortfall är ej kvantifierad men uppskattas ligga på kanske 5 % eller högre. Trafikolyckor med personskador (ca 1,6 % av alla viltolyckor) rapporteras vidare till databasen STRADA. Vår analys visade att i genomsnitt 37 % av alla viltrelate-rade personskadeolyckor under 2003–2012 inte var klassade som en viltolycka och därmed förmodligen inte heller fanns listat som viltolycka i Polisens T-RAR (trafikregister). När en olycka med vilt anmäls kontaktas i regel en kontaktperson kontaktperso-ner för att vid behov ta hand om det skadade djuret. Vid jägarens platsbesök skrivs en viltolycksrapport med detaljerade uppgifter om plats, tid och djurart. Inte alla rapporter (84 %) innehåller dock fullständiga och korrekta uppgifter som kan användas vid rumsliga analyser av olyckor. Det verkar även finnas en skevhet i viltolycksstatistiken med ca 12 % färre eftersök längs enskilda och kommunala vägar än förväntad. Tekniska problem i Polisens datasystem under 2012–2015 medförde dessutom att ett okänt antal anmälda viltolyckor inte registrerades i trafikregistren (T-rar, pust, k-rar). Det ledde till att det i vissa regioner och under vissa år finns fler viltolycksrapporter än polisrapporter. Vi uppskattar att detta bortfall kan ligga på minst 11 % i genomsnitt, under antagandet att rapporter från eftersöksjä-garna utfärdats för runt 90 % av alla polisregistrerade fall. Förenklat uttryckt står alltså den polisiära statistiken på viltolyckor under 2010 – 2015 för runt två-tredje delar av de verkligt inträffade olyckorna, medan efter-söksstatistiken i rumsliga analyser representerar ungefär hälften. Denna enkla tumregel bör dock användas försiktigt eftersom det finns tydliga avvikelser i mörkertalen mellan viltarterna, länen och åren. De olika bortfallen i statistiken har därmed en påtaglig effekt på den totala upp-skattningen av olyckstal. Vissa av orsakerna för bortfallen kan dock lätt åtgärdas eftersom de beror till stor del på brister i rutiner för registrering och data-bashantering. Vi rekommenderar därför en systematisk kontroll av databaserna och förbättrade kontrollmekanismer vid registrering och klassificering av an-mälda fall. Vi rekommenderar också att databaserna i STRADA, trafikregistret och eftersöksstatistiken länkas via ett gemensamt händelse-ID. Detta kan un-derlätta mycket vid framtida bortfallsanalyser. Vi föreslår även detaljstudier av hur olyckorna registreras och tolkas för att bättre kunna förebygga framtida bortfall. Hidden statistics in wildlife-vehicle collision data – results from a drivers’ questionnaire and database analyses Among the most common causes of road accidents in Sweden are collisions with wild animals, especially ungulates. Over 50,000 ungulate accidents per year have been reported during the past 5 years and the numbers are steadily in-creasing since the 1970-ies. Despite regular campaigns, extensive investment in wildlife fencing and other preventive measures, and in contrast to declining game bag in moose and roe deer, accident statistics increase faster than what can be expected from increased traffic alone. It is obvious that wildlife-vehicle accidents in Sweden are not under control. Major contributing factors are defi-ciencies in data and uncertainty in statistics. It is well known that not all accidents are reported or show up in the official statistics, but the correction factor still used by the Swedish Transport Administration is based on over 35 years old data when traffic and wildlife conditions had been different. A better knowledge of where and when and how frequent wildlife-vehicle collisions occur is needed to more effectively plan and target mitigation actions. This project provides updated estimates of the hidden statistics in wildlife-vehi-cle collision data and identifies uncertainties and problems in current statistics. This was done by: i. a survey with car drivers, ii. an analysis of wildlife-related accident statistics on human injuries in the Swedish Traffic Accident Data Ac-quisition (STRADA), and iii. a comparative analysis of accident report statistics from the police and from hunters provided by the National Council on Wildlife-Vehicle Collisions (Nationella Viltolycksrådet). To estimate how often drivers do report and refrain from reporting wildlife-ve-hicle accidents to the police, we conducted a publicly available online-survey during Nov. 2013 to Dec. 2014 that was answered by 3981 respondents. We asked the respondents about their knowledge of and experience with wildlife-ve-hicle accidents and requested details on the most recent incident they experi-enced after 2004. More than half of all respondents (65%) reported to have been involved in wildlife-vehicle collisions at any time and around 20 % of these indi-cated that the accident was never reported to the police. About 45% of the re-spondents declared that they experienced traffic accidents with wildlife after 2004, and the proportion of non-reported incidents was estimated to between 9% (public respondents) and 19 % (control groups). We recommend therefore assuming that about 15% (± 5%) of the incidents will not be known by the police. Compared to studies from the late 1970-ies, these hidden statistics appear hence much smaller today. After that an incident has been reported to the police, however, there are further important sources of data loss. Depending on how the incident has been classified in the report, the records are manually transferred to secondary databases such as the traffic accident register (T-RAR). Incidents where wildlife was not the immediate cause of the accident may not be classified as wildlife-vehicle collision and thus not be found in the of-ficial police statistics. This loss has not been quantified but is estimated to be around 2%. Traffic accidents with human injuries (about 1.6% of all wildlife accidents) are reported to the database STRADA. Our analysis showed that on average 37% of all game-related injury accidents during 2003- 2012 was not classified as a wild-life accident and therefore probably neither was listed as such in the official police statistics. In most cases (74%) when an accident with wildlife is reported to the police, they notify a contracted hunter to take care of the injured animal. The hunter in turn issues a report with detailed information on the location, time and animal species. Not all of the reports (84%), however, contain complete and accurate in-formation that can be used for spatial analyses of accidents. In addition, hunt-ers’ reports seem to be biased towards larger roads and underestimate the num-ber of accidents on private and tertiary roads by about 12%. Technical problems in the police data system during 2012-2015 resulted in the loss of an unknown number of reported wildlife-vehicle accidents. This is why in some regions and in some years, more hunter reports were issued than police records exist. We estimate that this loss may accede 11% on average. In simplified terms, police statistics on wildlife-vehicle accidents during 2010 - 2015 may stand for about two-thirds of the truly occurred accidents, while hunter reports that are used in spatial analyses represent about half of the true accident frequency. However, this rule of thumb should be used cautiously as there are substantial differences in the hidden and lost statistics between the years, counties and species. Thus, the various shortfalls in these statistics have a significant impact on the overall estimate of wildlife-vehicle collision numbers. Depending on the data sources and on how statistics are used, different biases and data losses must be considered. Some of the causes can be easily overcome as they relate to deficien-cies in registration routines and database management. We therefore recom-mend a systematic check of the databases and improved control during registra-tion and classification of reported cases. We advocate that the different inde-pendent databases are linked through a common event ID. We also suggest de-tailed studies of how accidents are recorded and interpreted in order to better prevent future data loss. Preliminär rapport, version 2017-05-31. Trafikverkets publikat-ionsnummer 2016:055

Traffic accidents with ungulates pose a serious problem in many countries, and there is a need for predicting accidents and costs at a large scale for an efficient management of the accidents. Based on the assumption that traffic accidents are determined by traffic volume and ungulate population sizes, this study provides a relatively simple method for calculating and predicting costs of current and future traffic accidents with roe deer, wild boar, and moose in Sweden. A logistic population model is assumed for all ungulates, and econometric methods are used for predicting vehicle accidents with panel data on traffic accidents, traffic load, bags, hunting licenses, and landscape characteristics for each county and year during 2003-2015. The calculated total discounted cost of traffic accidents over a period of 15 years is relatively stable around 1300 million SEK per year in present value (which corresponds to 0.03% of gross domestic product in 2015), but the allocation of costs among ungulates differs. Costs of vehicle accidents with moose account for the largest share of the cost (44%), but accidents with wild boar show the most rapid increase over a 15 year period because of the estimated relatively high intrinsic growth rate and the recent establishment of this animal in several counties. The predicted costs are, however, sensitive to the assumption of future hunting pressure and traffic volume.