Standardized monitoring of Rangifer health during International Polar Year

Monitoring of individual animal health indices in wildlife populations can be a powerful tool for evaluation of population health, detecting changes, and informing management decisions. Standardized monitoring allows robust comparisons within and across populations, and over time and vast geographic regions. As an International Polar Year Initiative, the CircumArctic Rangifer Monitoring and Assessment network established $eld protocols for standardized monitoring of caribou and reindeer (Rangifer tarandus) health, which included body condition, contaminants, and pathogen exposure and abundance. To facilitate use of the protocols, training sessions were held, additional resources were developed, and language was translated where needed. From March 2007 to September 2010, at least 1206 animals from 16 circumpolar herds were sampled in the $eld using the protocols. Four main levels of sampling were done and ranged from basic to comprehensive sampling. Possible sources of sampling error were noted by network members early in the process and protocols were modi$ed or supplemented with additional visual resources to improve clarity when needed. 'is is the $rst time that such broad and comprehensive circumpolar sampling of migratory caribou and wild reindeer, using standardized protocols covering both body condition and disease status, has been done.


Introduction
Understanding and tracking the health status of individual animals provides valuable information for wildlife management (Jean & Lamontagne, 2004;PCMB, 2010;GNWT, 2011).Changes in body condition or infectious disease indices of individuals may re ect shifts in population health and serve as early warning signals for wildlife managers, wildlife users, and public health o cials concerned with zoonotic diseases.Examining interrelationships between various health indicators, as well as their trends over time and across herds and geographic regions, provides new insights into the ecology of a species and functioning of an ecosystem.Foundational to these activities, , 2012 however, is securing a comprehensive baseline of the existing health status of the population (Karesh & Cook, 1995), including but not limited to: body condition, genetic diversity, physiological health, contaminants, and pathogen diversity and abundance.Understanding the relationships among various health indicators, and de ning what is 'normal' and the variability around that normal within a host population, through population cycles and across the range of a species, is an important rst step for identi cation of 'abnormal' and early detection of and response to changes in health.
Establishing baselines and maintaining monitoring programs for wildlife populations is not easy.Financial constraints, political and/ or philosophical di erences, transboundary issues, and the general elusive nature of wildlife make it di cult to establish and maintain successful programs (Ko nas et al., 2002;Witmer, 2005).In the Arctic, widely scattered communities, a vast landscape, high costs of eld access for research and hunting, and logistical constraints, add to the di culties of wildlife monitoring. is is particularly true for caribou and reindeer (Rangifer tarandus sspp.), a very mobile keystone species (Geist, 1998) that is highly valued for cultural and economic reasons across its circumpolar range (Nuttal et al., 2010).Nevertheless, a variety of Rangifer monitoring programs of di erent types have existed over time (Ko nas et al., 2002;Couturier et al., 2004;Gunn et al., 2005;Campbell, 2006;Campbell et al., 2010;Lyver & Nation, 2010).
Monitoring of a sort began with aboriginal people who, for thousands of years, have been keenly aware of the health of caribou or reindeer that they harvested for food, clothing, and tools (Cruikshank, 1981;Ferguson & Messier, 2010).A technical approach to monitoring body condition began in the mid-1900s, with emphasis on describing and validating condition indices which had allometric relations to the status of muscle, bone and fat, and their chemical constituents: water, protein, lipid, and ash (Dauphiné, 1976;Langvatn, 1977;Ringberg et al., 1981a;b;Reimers & Ringberg, 1983;Huot & Goudreault, 1985;Adamczewski et al., 1987a;Huot & Picard, 1988;Allaye Chan-McLeod et al., 1995;Gerhart et al., 1996).Monitoring body size (e.g., lower jaw and metatarsal bones) and shape (e.g., heart girth) also provides insight into population trends (Parker, 1981;Crête & Huot, 1993;Mahoney & Schaefer, 2002;Couturier et al., 2010).For the most part, this type of monitoring has been scientist-driven, typically initiated and conducted by wildlife managers or researchers, and often focused on a particular herd intermittently or for a nite length of time.
Opportunistic monitoring of infectious diseases of Rangifer by scientists in partnership with aboriginal hunters has also occurred since the mid-1900s, often initiated by hunter reports or submissions of abnormal tissues (e.g., Canadian Cooperative Wildlife Health Centre http://www.ccwhc.ca/;Choquette et al., 1967).More recently, targeted programs for contaminants and disease surveillance have developed across many jurisdictions (Elkin & Bethke, 1995;Zarnke et al., 2000;Robillard et al., 2002;Brook et al., 2009;Stieve et al., 2010).
Increasingly, through the wildlife co-management process, Rangifer users themselves are driving the context for, and implementation of, monitoring (Lyver & Gunn, 2004;Brook et al., 2009;Lyver & Nation, 2010).People who depend on Rangifer for food and income have expressed concerns about how this species, and those who depend on it, will cope with the increasing rate of environmental and political change (Ko nas et al., 2003;Brook et al., 2009).e impacts of climate change, resource development and other stressors on the health of Rangifer, and on food safety and security (i.e., population sustainability) as it relates to Rangifer, are major concerns for subsistence hunters (Brotton & Wall, 1997;Brook et al., 2009).
In light of global climate changes and conservation e orts, the Arctic Council launched a number of monitoring programs for Arctic species through the Conservation of Arctic Flora and Fauna Group (CAFF, 2010).e Circum-Arctic Rangifer Monitoring and Assessment (CARMA) network, positioned within CAFF, was launched in 2004 in response to increasing concern for Rangifer and the need for circumpolar collaboration.e network is a forum to exchange ideas, observations and data, and coordinate Rangifer monitoring activities around the Arctic.
A key objective of the CARMA network was to develop and implement standard methods for monitoring Rangifer health.Although monitoring activities were already in place for many herds, the methods, frequency, and type of data recorded varied greatly (Ko nas et al., 2002).Integration of data across disciplines (e.g., disease and body condition data), even within a herd, was sometimes lacking.Standardized approaches, both within and among herds, allow comparisons across space and time and, therefore, can provide a much deeper understanding of Rangifer health.As standardized baselines become established, links between the various health indicators can be examined and the costs, or bene ts, of pathogens and pollutants at the individual, population and community levels can be evaluated.Such information provides the necessary foundation to assess Rangifer vulnerabilities and responses to environmental and anthropogenic changes.Importantly, for standardized protocols to be adopted and e ective, the procedures need to be clear and without ambiguity, but exible enough to accommodate di erences in monitoring programs and objectives.At the same time, they must include enough indices to allow predictions: for example, to infer the probability of pregnancy from the fat and protein reserves (Ko nas et al., 2003).
Supported by the International Polar Year (IPY) initiative (2005 -2011), the CARMA network developed a standardized approach to circumpolar monitoring for Rangifer.In this paper we provide an overview of the CARMA protocols and the implementation of these protocols for CARMA-supported sampling of circumpolar caribou herds.We summarize the IPY sampling e orts, and discuss the successes and hurdles to such broad monitoring activities.

Standardized protocol development and application
Starting in 2006, a sub-group of the CARMA network developed: (i) a manual that described monitoring indicators, the rationale for each indicator, and the relevant literature, and (ii) standardized sampling protocols for collection and measurement of each indicator.e selection of indicators and sampling protocols were developed in consultation with network collaborators.ese were based on published and unpublished literature and experiences from previous and ongoing Rangifer monitoring programs.Indicators were selected to provide data on age, diseases, physiological condition, short and long-term nutritional status, and maternal investment in reproductive tness.Some indicators were compatible with, and could be used in energy-protein and body frame size models that CARMA developed to assess and predict Rangifer responses to environmental changes (Murphy et al., 2000;Russell et al., 2005).
Recognizing the value and constraints of both hunter and scientist-based monitoring, CARMA initially developed two levels of protocols.e Level 1 protocol was developed for subsistence hunters or community-based sampling and provided basic information on age, frame size, body condition, and a few pathogens.
e Level 2 protocol was more com-prehensive and typically required at least one well-trained individual to collect and process samples.Additionally, this level required partnerships with commercial or research laboratories to do further analyses on samples.Level 2 included additional indicators for body condition and morphometrics that could be used to quantitatively predict whole body fat and protein reserves, and in-depth sampling methods for assessing contaminants and determining pathogen presence, abundance, and/or exposure.
In the protocols, the sampling procedures and data requirements for each indicator were described and references were provided for further details on more specialized procedures such as tooth sectioning or parasite isolation from tissues.
e manual and protocols included less detail on pathogens and pathogen sampling than was applied in the CARMAsupported eld collections so these procedures are described in greater detail in this paper.
CARMA network partners and participants were encouraged to follow the protocols when handling Rangifer during IPY-funded collections.e initial selection of herds and the level of sampling was discussed at the annual CAR-MA meetings and aimed to identify representative reference herds around the Arctic (Fig. 1).
e nal selection of the reference herds was determined by the priorities and support of management agencies, communities, and comanagement boards.Further, the sampling intensity for each individual herd was in uenced by the availability and engagement of regional biologists or managers and by the goals of individual research and monitoring projects.
Formal feedback on protocol implementation was gathered during the 2010 annual CARMA workshop after network members had an opportunity to use the protocols in eld collections from 2007 -2009.Network members (63 people representing 11 government agencies, 12 organizations, eight universities, and six countries) divided into three groups (community people, wildlife managers, and university researchers) and had guided open discussions on whether the protocols were useful and how content and accessibility could be improved.

Data management
All participating herd biologists or managers were asked to submit their data in a Microsoft Excel® datasheet format and all entries were imported into a newly created CARMA Excel datasheet.When possible, an automatic importation function (e.g., "vlookup") was used to decrease human error.When needed, original datasheets were consulted to validate the data.In the absence of eld datasheets, herd biologists or managers were asked to verify and validate their data once it was imported into the common database.For most collections, metadata records, documenting the date, location, number of animals, and types of data collected, were created by individual researchers and archived with the ArcticNet data portal (www.arcticnet.ulaval.ca).

Standardized protocol development
Draft protocols were discussed with network members before and during IPY at the CAR-MA annual meetings as well as between meetings as needed (e.g., manager/researcher requiring speci c advice on sampling).Protocols were re ned periodically based on these discussions and ongoing feedback and in some cases, based on the results of research studies.For example, the initial protocol for Besnoitia included sampling multiple tissues, but based on results from this sampling it was determined that metatarsal skin alone was a suitable index of infection (Ducrocq et al., 2012).e manual and eld protocols are available on the CARMA web site (CARMA, 2011a;b;c).An early version of the protocols was also translated into Russian to be tested by CARMA's Russian collaborators.Indicators and samples collected, the information gained, and some of the potential sources of error associated with collections are summarized in Table 1.

Field collections and application of the protocols
Between March 2007 and September 2010, CARMA collaborators collected body condition and health samples from at least 1206 individual caribou and reindeer from 12 North American, two Greenlandic, and two Russian herds (Fig. 1; Table 2).e Chukotka herd was also sampled as part of the IPY initiative but data from this herd were not available at the time of writing.
Sampling intensity for each herd varied depending on the objectives of the monitoring project.Although only two levels of protocols were initially designed (see methods), in practice, sampling could be categorized into four overlapping, categories.

Level 1 (n = 152 animals)
is was the simplest type of sampling, required the least amount of training and data recording, and was generally done by subsistence hunters.Often hunters were paid for samples and data submission.Hunters were asked to collect information on a minimum number of indicators that were selected to provide important basic data while interfering minimally with meat handling practices (Table 1).Hunters lled in tags (Fig. 2a) to identify , 2012    (Parker, 1981; Fig. 3a) was measured using the provided tags or rulers.Samples submitted included the lower jaw and metatarsal bone (Fig. 3b) with skin on, and core data collected included age (using tooth cementum annuli) (Miller, 1974;Matson, 1981), pregnancy, body condition (back fat, marrow fat, and a subjective hunter assessment; Lyver & Gunn, 2004), body size (jaw and metatarsal length; Parker, 1981;Couturier et al., 2010), and presence and intensity of the parasite Besnoitia tarandi in metatarsal skin (Ducrocq et al., 2012).Besnoitia cysts were quanti ed in the lab on gross observation using categories of number of cysts per square centimeter [0 = no cysts; 1 = very occasional (< 4 cysts/cm²); 2 = easily observed (4-10 cysts/ cm²); 3 = high density (> 10 cysts/cm²)].Histological examination, however, was determined to be much more reliable both for presence and intensity (Ducrocq et al., 2012).Level 1 provided key samples and data that could be analyzed to evaluate and compare body condition and size, demographics (recognizing that hunter-based sampling may have strong, but identi able biases), and Besnoitia infection status.
Level 2 sampling (n = 166 animals) is type of sampling was typically done by trained subsistence hunters or during community organized hunts that were attended by biologists and veterinarians.Samples included those from Level I as well as blood collected on lter paper (Fig. 3c; Curry et al., 2011), the left kidney with fat, liver, and feces (Table 1).Semiquantitative observations were done for grossly visible parasites and, depending on the collection, some body measurements were recorded (Fig. 2b; CARMA, 2013).
Collection of blood samples allowed for more in-depth examination of pathogen diver- sity (Curry et al., 2011).Blood was extracted from the lter paper in the lab and tested for exposure to a variety of pathogens using standard serological tests (Curry et al., 2011).Additional testing for Trypanosoma spp.and Setaria sp. was attempted using polymerase chain reaction-based approaches.DNA was extracted from lter papers using a modi ed PurGene® (Qiagen, Canada) extraction procedure for compromised blood samples.Primers described by Noyes et al. (1999) were used to test for the presence of Trypanosoma and the primers described by Laaksonen et al. (2009) were used to test for the presence of Setaria.Unfortunately, DNA quality was variable and DNA was not reliably extracted from the lter papers (D.Schock & S. Kutz, unpubl. data).
us, although the presence of Trypanosoma and Setaria was con rmed from some samples (Kutz et al., 2012), the absence of these parasites could not be con dently established and quantitative assessment was not possible.It was, however, possible to sequence several Trypanosoma detected in caribou samples and compare genotypes among caribou herds and among ungulate species (D.Schock & S. Kutz, unpubl. data).
e kidney with surrounding fat was collected and used to establish the Riney kidney fat index (Riney, 1955).e kidney, together with a piece of liver, usually the caudate lobe, was frozen for future analyses.Feces were stored frozen and later tested for helminth and protozoan parasites and the bacteria Mycobacterium avium paratuberculosis (Forde et al., 2012).
is dataset provided substantially more quantitative and semi-quantitative information on body condition and pathogen presence, exposure, and abundance than Level 1.It therefore allowed for more extensive investigations of pathogen/parasite occurrence in relation to body condition and other physiological and demographic parameters.

Level 2 detailed sampling (n = 835 animals)
ese were the most comprehensive collections.ey were planned hunts done primarily for scienti c purposes and/or management monitoring, and in some cases were done in collaboration with community hunts.Typically, these collections involved local hunters and the meat was given to the local communities.In addition to the data and samples listed above, more detailed body size measurements and indepth pathogen assessments were done (Table 1; CARMA, 2013).Project speci c sampling was done during these collections, varied across herds, and samples may have included sampling: the brain stem ventral to the obex and/or lymph nodes for chronic wasting disease (Williams, 2005), conjunctiva and skin from the rostrum, scrotum, and inner thigh for Besnoitia research (Ducrocq et al., 2012), abomasum and rst three meters of small intestine for gastrointestinal parasite analyses, a 5-10cm section of , 2012 ileum and mesenteric lymph nodes for Mycobacterium avium paratuberculosis, ovaries (Dauphiné, 1978), various tissue sections in formalin for describing normal histology, weight of the rumen contents (Huot, 1989), and additional body measurements.Whole (EDTA tubes) and clotted blood for serum were often collected in addition to blood on lter paper.Lungs were dissected along the bronchi and major bronchioles to assess Dictyocaulus infection (Anderson & Prestwood, 1981) and livers were cut into 1 cm slices to examine for the liver uke, Fascioloides magna, and Taenia hydatigena cysts (Lankester & Luttich, 1988).Transverse cuts through both the heart and the gastrocnemius muscle, and sometimes additional muscles, were done to examine for cysts of Taenia spp.Feces, muscle, fur, and urine were also collected for protein/nitrogen balance studies (Barboza & Parker, 2006).
is sampling level allowed for more detailed quantitative analyses of a broader range of pathogens.e broader dataset on body condition indices (back fat, kidney fat, marrow fat, empty and dressed body weight, and cranial crural muscles) provided the data necessary to predict body condition, fat, and protein (Ring-

Live animal sampling (n = 53 animals)
is sampling was done by biologists as a routine component of animal handling during radio-collaring.Data collected included an estimate of age (based on body size and tooth wear), body condition, feces, blood, and hair.Depending on season these animals were allotted in reproductive category (pregnant/nonpregnant, lactating/non-lactating, and in some cases weaning status).Body condition and reproduction status could also be examined relative to status of pathogens that could be grossly observed (e.g., Besnoitia), or those that could be assessed serologically or through fecal examination.

CARMA network feedback and protocol evaluation
Overall, the protocols were well received and consistently applied.Formal feedback from the network, gathered during three breakout groups in 2010, was positive with helpful comments to improve the protocols and their utility (CARMA, 2013).Network members emphasized the importance of the written protocols and visual resources for sampling and laboratory processing (images and directions in the protocols, sampling video, anatomy website), and suggested that these be translated into the languages of the various user groups around the Arctic.Aboriginal members of the network indicated that their view and observations of Rangifer 'health' may di er from the scienti c perspectives and that a community-developed protocol that incorporated this view for health monitoring would be valuable.Additional suggestions included: development of advanced protocols for non-lethal sampling, improved guidance on necessary sample sizes for power analysis, and additional guidance on interpretation of results.
Network members did identify some issues with interpretation of the eld protocols (Table 1).One concern was identifying the anatomical limits of the 'peroneus' muscle, the weight of which could be used to estimate body pro- s tein. is is actually a complex of four muscles, better referred to as the cranial crural muscles.ey are found on the front (cranial aspect) of the shinbone (tibia) in the crus or crural part of the leg; hence the name.ey form a coherent functional group that, taken together, ex the hock and extend the digit.e cranial crural muscles are relatively brous when compared with other muscles and are not much sought after for food.
ey can be quickly and consistently removed as a group, and weighed without further dissection.To clarify the muscle group to be sampled, a series of images were produced by dissections of reindeer and caribou by Peter Flood, Christoph Muelling, and others (Fig. 4a, b).A step-by-step instructional Microsoft PowerPoint® presentation describing the appropriate anatomy and sampling process for this muscle group was also produced and is available at http://www.ucalgary.ca/caribou/Sampling.html).
Other potential sources of error were identied when datasets were merged into a single database and it became clear through this process that, to avoid errors, complete descriptions of any modi cations from the standard protocols needed to be included with all metadata les.For example, the total body length measured for some herds included the tail whereas the tail was excluded for others.Occasionally confusion arose over the units of measurement.In some cases these were not speci ed on datasheets.For example, back fat could be measured as 1 mm versus 1 cm, or 1 inch versus 1 cm.
is was particularly an issue if measurement tapes included both imperial and metric units.Some datasets contained blank spaces that left ambiguity as to whether an animal had been examined for that speci c indicator/pathogen and zero observed, or if the indicator had not been examined at all.Errors, blanks, or unspeci ed deviations from protocols were recti ed by clari cation from contributors.
Sampling and data collection for pathogens were generally standardized and consistent, however, some clari cations to the protocols were also needed to prevent loss of data.For example, quanti cation for serology requires that lter paper strips are fully saturated.In some cases the blood strips were only partially saturated and results were not quanti able or comparable.Quanti cation of Taenia cysticerci di ered between protocols.Speci cally, at level 2, the hunters were only asked if they saw any Taenia cysts in the carcass.In contrast, in the detailed level 2 protocol, two horizontal cuts were made through the gastrocnemius and the heart, and more intensive sampling (e.g., more cuts or organs examined) was done for some herds.Once samples for pathogens were collected, standardization in laboratory analyses was achieved by ensuring similar storage of samples among herds and the use of the same laboratory for each pathogen.is limited sources of error and variability among herds.
In response to network member feedback and queries, early in the process two products were developed to supplement the protocols and provide additional visual resources.A DVD on basic and advanced sampling was produced in cooperation with hunters from the communities of Fort Good Hope and Colville Lake, Northwest Territories, Canada (CARMA, 2009).
e video was distributed to CARMA network collaborators, participating communities, as well as local and regional wildlife groups and agencies.It is available on the CARMA website (CARMA, 2013).In addition, the Rangifer Anatomy Project was initiated to produce high quality anatomical images of Rangifer and instructional PowerPoint presentations to supplement the protocols and serve as teaching aids for a variety of audiences.
e site is located at the University of Calgary (http://www.ucalgary.ca/caribou/index.html), and can be accessed through the CARMA website.
, 2012 Discussion CARMA's goal was to bring together the vast knowledge and expertise in the circumarctic so that sampling protocols incorporating local knowledge and science could be developed and provide reliable and comparable information on health and body condition of Rangifer across its range.is goal was achieved in that a series of standardized protocols with increasing levels of complexity were developed and implemented in varying degrees across herds.However, as evidenced by feedback in 2010, the CARMA protocols did not adequately incorporate the aboriginal views on health.Protocol development and implementation was a learning process, with challenges and limitations being identi ed and addressed on an ongoing basis.An advantage of CARMA's web-based approach was that it was responsive to identied needs.e protocols and manual were easily updated and clari ed and supplementary resources could be made readily available on the web. is process is ongoing.
An innovative aspect was the partnership with infectious disease specialists, which brought a strong emphasis on monitoring techniques for pathogens and non-infectious diseases.is emphasis was in recognition of two points.Firstly, body condition and disease monitoring has not always been well integrated in the past.Body condition and disease are intricately linked and neither can be well understood in isolation from the other.Secondly, the emergence of new diseases, re-emergence of old, and pathogen spill-over among wildlife, domestic animals, and people has highlighted the need for e ective and responsive wildlife disease monitoring and surveillance systems today (Kuiken et al., 2005).
It became clear during IPY that when dealing with a species that is widespread across vast geographical remote regions there is a need to develop exible yet scienti cally robust strategies for data collection.Although the protocols were originally developed as two de ned categories of collection, they were adapted to meet the monitoring, logistical, research, and human resource constraints for each herd. is highlights the importance of providing a prioritized continuum that allows useful and comparable information to be gathered even at the most basic level.A frequent limitation to monitoring is inadequate sample size, which can limit statistical power to detect trends in concomitant variables (Nickerson & Brunell, 1998).Importantly, "required" sample size will vary for each speci c pathogen depending on its expected prevalence.Several authors emphasize that working with hunters to monitor health and condition of harvested caribou increases sample sizes (Ko nas et al., 2003;Lyver and Gunn, 2004;Brook et al., 2009;Curry, 2010).e protocols accommodate the trade-o between the greater sample sizes available from hunterbased collections compared to the more detailed dataset from fewer animals collected during intensive monitoring and research projects.For example, a level 1 collection (i.e., the jaw, metatarsal, and other basic data) can provide reliable information on late term pregnancy, age and sex structure, body size, body condition, and abundance of a few select pathogens.Additional information on other physiological parameters can also be gathered from hair (e.g., cortisol and other hormones; Ashley et al., 2011) while teeth can be examined for dental enamel hypoplasia as an indicator of past stress events (Wu et al., 2012).Hunter-based sampling was done prior to IPY across a number of jurisdictions and has, in general, broad acceptance amongst hunters (Gunn et al., 2005;Brook et al., 2009).With widespread hunter involvement, this type of monitoring could provide ongoing, reliable and a ordable information on several indices of Rangifer health across a broad geographic range and across seasons (Ko nas et al., 2003).As the complexity of data and sample collection increases (e.g., through assisted community hunts and biologist led collections) more information is acquired and more comparisons are possible.However, such comprehensive monitoring is typically done on fewer herds and over shorter time frames, limiting the number of herds or years that can be compared.us, establishing a 'bare minimum' guideline for sample and data collection facilitates ongoing broader (geographically and temporally) comparisons at a basic level.Importantly, all sampling strategies have biases speci c to that strategy, for example, subsistence hunters may select for animals in better condition, and these must be considered in the nal data analyses.
e additional spatiotemporal sampling and increased sample sizes o ered by hunter-based sampling may be particularly valuable for disease surveillance as it can increase detection of infectious diseases that may be present at a low prevalence and/or with a patchy or clumped distribution (Zhang et al., 2011).Working with hunters, however, o ers important advantages beyond simply increasing sample size.
ese include two-way exchange of knowledge, recognition of di erent ways to assess health and condition, and growing mutual respect and trust developed from working together (Konas et al., 2002;Brook et al., 2009;Lyver & Nation, 2010).
Emerging diseases are of increasing global importance, and are also of considerable concern for northern ungulates (Kuiken et al., 2003;Kutz et al., 2004;Laaksonen et al., 2010).Northward range expansion of domestic and wild animals, together with various pathogen vectors, may allow introduction of new pathogens into Rangifer range.Concurrently, a warming climate is removing some of the environmental constraints on existing arctic pathogens, resulting in range expansion of pathogens and emergence of disease (Hoberg et al., 2008;Kutz et al., 2009;Laaksonen et al., 2010).Establishing comprehensive baselines for pathogen biodiversity and abundance is necessary if we are to detect emergence of new pathogens and disease syndromes.Sophisticated molecular diagnostic techniques in combination with the advanced level of collection described here allows, through either physical recovery of the organism or serological evidence of exposure, detection of all known pathogens of Rangifer and the quanti cation/semi-quanti cation of most.
Monitoring programs must ensure accuracy and precision (i.e., repeatability) of measurements.e CARMA network approached this through the detailed descriptions in the manual and protocols, hosting speci c training sessions, and having biologists participate in collections before running collections of their own.Use of the same diagnostic and research laboratories in many cases helped to ensure standardization.
Preservation of biological specimens and data are critical elements of any monitoring program.In particular, appropriately preserved physical specimens allow for investigation of new questions, or re-evaluation of old studies, as new information and techniques emerge (Hoberg et al., 2008).Archiving of CARMA tissues was limited to the capacity of individual researchers and, for those speci c caribou populations, provides a rich source of information.Importantly, to ensure appropriate preservation of these materials, and to promote increased rates of archiving in the future, centralized, permanently curated facilities for specimen and DNA archiving are essential.
Data management in large-scale monitoring programs adds complexity at several levels.Data ownership and management is an important issue within any large network.Metadata for CARMA were managed through the Arctic-Net portal, however, mechanisms for long-term storage of full datasets in a central database and subsequent access remains to be established.all data, tissues, and specimens should be made available to the broader scienti c community after a reasonable time period.Sharing data also raises questions of format and transcription errors when inputting from data sheets to a database.Exploring the use of a computer program for data mining would reduce transcription errors while transferring data from original databases.CARMA continues to discuss the development and implementation of solutions.

Conclusion
CARMA is a network of Rangifer users, biologists, scientists, and managers who have worked together to implement the rst broad scale standardized body condition and health assessment of migratory caribou.To our knowledge, this is the most comprehensive sampling e ort for a terrestrial mammal across its entire range.Network members were actively and willingly engaged in improving and implementing protocols, and although some challenges were encountered, network members philosophically supported the protocols and applied them.
is was a clear re ection of the shared vision of how a comparative approach across space (herds) and time (seasons and years) can substantially improve our current scienti c knowledge of Rangifer and enables us to monitor and detect changes more rapidly and con dently.
Flexibility and adaptability of sampling regimes are essential to ensure that speci c research and monitoring objectives can be addressed.Still, there was consensus that a minimum of standard data and samples should be done for all herds (e.g., Level 1).Easily accessible and up to date protocols provided key resources for eld and laboratory personnel.Where needed, these were supplemented with additional visual aids such as the anatomy resources and sampling DVD.Ongoing in-person and practical training and mentorship of hunters and researchers are essential to ensure accurate and consistent sample and data collec-tion into the future.Engagement of communities and local hunters is critical to promote knowledge sharing and mutually acceptable approaches for long-term health and condition monitoring in caribou.
Data management and ownership, together with authorship on scienti c articles, had the potential to become problematic.However, a transparent approach and regular in-person discussions enabled the growth and maintenance of productive collaborations among network members.To be e ective, the data and knowledge gained must be transferred back to communities and caribou managers in a timely and e ective manner.is has been done directly within the CARMA network as community representatives are key and active network participants, and has also occurred in many regions in the form of co-management meetings.Additional researcher-initiated posters and community meetings presenting preliminary results have occurred.Unfortunately, as the funding from IPY has ended, there are severe nancial constraints to returning to communities in person to present nal results, many of which will be coming out over the next several years.
Rangifer populations co-exist with increasing human abundance around the Arctic, and like many parts of the world, anthropogenic modications in parallel with natural phenomenon are in uencing their health and sustainability (Balmford et al., 2003).To understand these complex systems, wildlife research needs to focus on large-scale monitoring activities (Pollock et al., 2002;Balmford & Bond, 2005).
e CARMA network and the standardized sampling protocols that it has developed, will hopefully enable a more comprehensive understanding of migratory Rangifer populations, and provide new insights into the resilience of these animals under the current regime of environmental, social, and political change.
in the making of the sampling protocol DVD.
e participation of the following governments, management boards, and corporations was also very valuable: Tlicho Government, Wek 'eezhii Renewable Resources Board, Vuntut Gwich'in First Nation, Porcupine Caribou Management Board, Gwich'in

Figure 2 b
Figure 2 b

Fig. 3
Fig. 3 Examples of samples taken and data acquired from caribou sampling activities.(a) Back fat depth is measured at a 45 o angle from the base of the tail (see inset) with a measuring tape, a ruler, or the tag (see Fig. 2a).Backfat is measured at the deepest point.(Image credit: Wendy Nixon).(b) Metatarsal bone length measured by calipers (Image credit: Wendy Nixon).(c) Drying of Nobuto® lter paper strips that have been dipped in clean blood (Image credit: Karin Orsel).

Fig. 4 e
Fig.4e cranial crural muscle group of a caribou leg.(a) Lateral view of the hind leg showing the anatomical position of the cranial crural muscle group (spotlit).e cranial crural muscle group is composed of four muscles: a. the cranial tibial (tibialis cranialis), b. the long bular ( bularis longus), and c. the fused third bular ( bularis tertius) and long extensor (extensor digitorum longus).When consulting standard veterinary anatomy texts, the reader should be aware that bularis and peroneus are synonyms and that both have been widely used; it has recently been agreed that bularis should be the standard term (b) e cranial crural muscles separated.e long extensor and the third bular are extensively fused at their upper ends and the two tendons of the long extensor are visible behind the third bular.ere is no need to separate these muscles in the eld as they should all be weighed together.(Image credits: Rangifer Anatomy Project, P. Flood and C. Muelling).

Table 1 .
Levels of sampling done by CARMA with the samples and data gathered (indicators) at each level, the information that each indicator provides on animal body condition, disease, or contaminants, and observed or anticipated sources of error or sample quality compromise associated with collection or transport of the data or 1 Assessment of condition (i.e., 'good' in spring has less backfat then 'good' in fall).This journal is published under the terms of the Creative Commons Attribution 3.0 Unported License Editor in Chief: Birgitta Åhman, Technical Editor Eva Wiklund and Graphic Design: Bertil Larsson, www.rangifer.no, 2012 1 (Loison et al. ; 2 (Wu et al 3 ; 4 (Handeland et al. ; 5 (Haley et al. ; 6 (Elkin & Bethke, 1995; Robillard et al.

Table 2 .
Collections done by the CARMA network during International Polar Year activities, 2007 -2010.e types of collections, data, samples, and the number of animals from each herd that were sampled (n) are indicated.Blank cells indicate that no animals were sampled.
Russian collections had potentially an alternate measure of kidney fat that has yet to be calibrated to the KFI in the CARMA protocols; 3 From metatarsal skin sections; 4 From ileum and ileo-caecal lymph nodes, and fecal and serum samples; 5 From blood sampled by lter papers or serum samples; 6 A combination of any of the following heavy metals: Aluminum, Antimony, Arsenic,Barium, Beryllium, Bismuth, Cadmium, Calcium, Cesium, Chromium,  Cobalt, Copper, Gallium, Iron, Lanthanum, Lead, Lithium, Magnesium, Manganese, Mercury, Molybdenum, Nickel,  Palladium, Platinum, Potassium, Rubidium, Selenium, Silver, Strontium, allium, Tin, Uranium, Vanadium, and  Zinc; 7Per uorooctane Sulfonate.
To date, CARMA has drafted a data policy as well as data submission and request forms.Ideally, in the spirit of open access that IPY promoted, Editor in Chief: Birgitta Åhman, Technical Editor Eva Wiklund and Graphic Design: Bertil Larsson, www.rangifer.no