Status of two West Greenland caribou populations 2005 1) Akia-Maniitsoq 2) Kangerlussuaq-Sisimiut. Technical Report No. 61

Status of two West Greenland Caribou populations 2005 1) Akia-Maniitsoq 2) Kangerlussuaq-Sisimiut Technical Report No. 61, 2005 Greenland Institute of Natural Resources 2 Title: Status of two West Greenland caribou populations 2005 1) Akia-Maniitsoq 2) Kangerlussuaq-Sisimiut Authors: Christine Cuyler, Michael Rosing, Johannes Egede, Rink Heinrich & Hans Mølgaard Funding: Greenland Institute of Natural Resources Series: Technical Report No. 61, 2005 Publisher: Greenland Institute of Natural Resources Cover photo: Christine Cuyler ISBN: 87-91214-15-7 ISSN: 1397-3657 Reference: Cuyler, C., Rosing, M., Egede, J., Heinrich, R. & Mølgaard, H. 2005. Status of two West Greenland caribou populations 2005; 1) Akia-Maniitsoq, 2) Kangerlussuaq-Sisimiut. Greenland Institute of Natural Resources. Technical Report No. 61. Part I-II, 64+44 pp. Available from: The report is only available in electronic format. You can download a PDF-file of the report at this homepage http://www.natur.gl/publikationer/tekniske rapporter It is possible to achieve a print of the report here: Greenland Institute of Natural Resources P.O. Box 570 DK-3900 Nuuk Greenland Phone: +299 36 12 00 Fax: +299 36 12 [email protected] www.natur.gl 3 Status of two West Greenland Caribou populations 2005 1) Akia-Maniitsoq 2) Kangerlussuaq-Sisimiut By Christine Cuyler1, Michael Rosing1, Johannes Egede2, Rink Heinrich3 & Hans Mølgaard4 1Greenland Institute of Natural Resources, P.O. Box 570, DK–3900 Nuuk, Greenland 2Innannguaq 13-B, DK–3900, Nuuk, Greenland 3Nukalloq 11, B-205, DK–3905, Nuussuaq, Greenland 4P.O. Box 122, DK–3911 Sisimiut, Greenland; [email protected] Technical Report No. 61, 2005 Greenland Institute of Natural Resources 4 Table of Contents Summary............................................................................................ 5 Introduction ...................................................................................... 9 Methods ........................................................................................... 12 Results.............................................................................................. 17 Discussion ....................................................................................... 21 Acknowledgements......................................................................... 28 Literature cited ............................................................................... 30 Appendices 1. Region stratification & transect allocation Page 32. 2. Survey field method & statistical design Page 35. 3. Increasing the accuracy of aerial counts of caribou in western Greenland Page 37. 4. Aerial survey 2005 Akia-Maniitsoq caribou, Central region, West Greenland Page 45. 5. Aerial survey 2005 Kangerlussuaq-Sisimiut caribou, North region, West Greenland Page 50. 6. Recommendations for future Page 55. 7. List of terms Page 58. 8. Comments from local observers Page 60 Part: II 9. Transect snow cover and visibility conditions. Photographs taken during the helicopter survey, March 2005. A) Central region, 14-16 March 2005 Page 1. B) North region, 18-22 March 2005 Page 14. 10. Caribou trails in the North Region Page 38. 11. Loess storms near the Ice Cap in the North region Page 41. 12. “Invisibility” of muskoxen & caribou and living side-by-side; Photographs taken March 2005 Page 42. 5 Summary In March 2005 two stocks, Kangerlussuaq-Sisimiut and Akia-Maniitsoq were surveyed by helicopter for abundance and herd structure. Methods and analysis followed Cuyler et al. (2003). Recommended stocking densities are exceeded by ca 60,000 caribou in the Kangerlussuaq-Sisimiut population and ca 17,500 in the AkiaManiitsoq. The 2005 calf percentage and recruitment are low. There is also a decrease in the ratio of bulls to cows. The low production and the skewed sex ratio may become more pronounced in future. Hunting pressure is not suspected to be a major cause for these changes, because the low harvest numbers can have had little effect on these large populations. Although quantitative data is lacking, it seems that density dependent effects are causing higher mortality rates among calves, and perhaps even among bulls, in the Kangerlussuaq-Sisimiut and Akia-Maniitsoq populations. Given that densities are three to six-times the recommended target value considered sustainable, we expect strong competition between individuals for available food resources. If the stocks are allowed to continue at their current size or increase further, there is a clear risk of lasting damage to the ranges, e.g. overgrazing and trampling. Unsustainable range use may compromise the future health and viability of caribou stocks in West Greenland. Regardless of management initiatives taken now, population crashes may be inevitable for some West Greenland herds within the foreseeable future, but accurate predictions about herd trends are impossible. To understand approaching developments the caribou and their range must be studied within the wider context of global warming and associated climate change. Akia-Maniitsoq herd – Central region The estimate for pre-calving population size of Akia-Maniitsoq herd of the Central region in March 2005 is ca 35,807 caribou (27,474 - 44,720; 90% CI). Caribou density in 2005 was 3.0 caribou per km2 in the high-density stratum, and 1.1 per km2 in the low-density stratum. Mean group size was 4.33 ± 2.91 S.D. in 2005. Late winter calf percentage was a low 14%, as was the annual recruitment of 24 calves per 100 cows. The bull to cow ratio was only 0.45. If natural mortality is between 8 and 10% then on a herd this size between 2,200 and 4,500 animals may be expected to die annually of natural causes. Kangerlussuaq-Sisimiut herd – North region The estimate for pre-calving population size of Kangerlussuaq-Sisimiut herd of the North region in March 2005 is ca 90,464 caribou (70,276 – 113,614; 90% CI). Caribou density in 2005 was 6.2 caribou per km2 in the high-density stratum, and 2.3 per km2 6 in the low-density stratum. Mean group size was 4.63 ± 3.38 S.D. in 2005. Late winter calf percentage was a low 11%, as was the annual recruitment of 16 calves per 100 cows. The bull to cow ratio was only 0.33. If natural mortality is between 8 and 10% then on a herd this size between 5,600 and 11,400 animals may be expected to die annually of natural causes. Since the calf percentage is approaching the natural mortality value, this population may be approaching its theoretical carrying capacity where births equal deaths, i.e. production equals zero. Eqikkaaneq Tuttoqarfinni Kangerlussuaq-Sisimiut kiisalu Akia-Maniitsoq qulimiguulik atorlugu marts 2005-imi kisitsisoqarpoq. Kisitsineq iluatsillugu tuttoqarfinni taakkunani tuttut eqimassusiat tuttullu ataatsimoortut qanoq aaqqissuunneqartarsimanerat ilanngullugu misissuiffigineqarpoq. Kisitseriaaseq misissueriaatsillu 2003-mi Cuyler allallu allaaserisimasaat malillugit ingerlanneqarput. Tuttut eqimassusissaannut innersuussutit naapertorlugit tuttoqarfimmi Kangerlussuaq-Sisimiut-mi tuttut 60.000-t missaannik amerlavallaalersimapput, kiisalu tuttoqarfimmi AkiaManiitsumi 17.500-t missaannik amerlavallaalersimallutik. 2005-imi kisitsinermi tuttoqatigiit tamakkerlugit isiginiarlugit piaqqat amerlassusiat ikigisassaavoq. Kisitsinermi aamma maluginiagassaavoq pannerit amerlassusiat kulavaat norraatsullu amerlassusiannut sanilliullugit ikiliartulersimanerat. Piaqqiaasartut ikiliartulernerat aammalu tuttut amerlassusiat eqqarsaatigalugu suiaassutsimikkut kipunganerat siunissami malunnaateqarnerulersinnaavoq. Allannguutinut pineqartunut piniagaanerinut tunngassuteqartut peqqutaasimanavianngillat, tassami tuttoqarfinni taakkunani tuttorpassuaqarnera eqqarsaatigalugu pisaasartumerngit annertunerusumik sunniuteqarsimanavianngimmata. Naak paasissutissat suli amerlanerusut pisariaqartikkaluarlugit tamaattoq ilimagaarput norraat allaallu immaqa pannerit akornanni toquinnartartunik naammattuuisarnerit tuttut amerlavallaarnerinik peqquteqarsimassasoq. Piujuaannartitsinissaq anguniarlugu tuttut amerlassusissaattut innersuussutinut sanilliullugu maanna tuttut pingasoriaammiit allaat arfinileriaat angullugu amerlanerulersimapput, taamaattoqarneralu peqqutaalluni naatsorsuutigineqartariaqarpoq nerisassarsiorlutik sakkortuumik unammisarsimanissaat. Tuttut maannatuut amerlatigiinnassappata imaluunniit amerliartuinnavissappata nunap neriniarfiusartup innarlerneqarsinnaanissaa ilimanaateqarluinnarpoq, tassa naggorlutsilluni imaluunniit tummaatarineqarpallaarnermigut aserorneqarluni. Taama piujuaannartitsinissamik tunngaveqanngitsumik neriniarfinnik atuititsineq ingerlaannassappat Kalaallit Nunaata kitaani tuttut peqqissusiat aammalumi uumaniaannarsinnaanissaat ajornartorsiortinneqaleraluttuinnassaaq. Tuttoqassuseq 7 mianeriniarlugu aqutsinikkut periusissat suulluunniit maannakkut atuutitinneqaleraluarpata Kalaallit Nunaata kitaani tuttoqarfiit ilaanni siunissami qaninnerusumi tuttut tassanngaannartumik ikileriarujussuaratarsinnaanissaannik periarfissaq pinngitsoortinneqarsinnaagunanngilaq. Taamaakkaluartorli tuttoqassutsip siunissami qanoq allannguuteqarsinnaanissaa siorngeruminaatsorujussuuvoq. Tuttoqassutsip siunissami allannguuteqarsinnaanissaa paasisaqarfigerusunneqarpat maannamut misissugaasartunut sanilliullugu misissuinernik siamasinnerusunik, tassa neriniarfiinut tunngassuteqartunik misissuinernik aammalu silaannaap kissatsikkiartornerata kingunerinik misissuinernik, ilaqartinneqartariaqarput. Tuttoqarfik Akia-Maniitsoq – Qeqqa Kalaallit Nunaata tuttoqarfittut immikkoortua Qeqqa, tuttoqarfik Akia-Maniitsoq, martsimi 2005 tuttut piaqqiulinnginnerisa nalaani amerlassutsimikkut missingiunneqarput 35.807inut (27.474 – 44.720; 90 % KI). Tamaani tuttut amerlanerpaaffiini kvadratkilometerimut amerlassusiat 3,0-imiippoq, ikinnerpaaffiinilu 1,1-imiilluni. Agguaqatigiissillugu ataatsimoortut 2005-imi amerlassusiat 4,33±2,91 SD-miippoq. 2005-imi ukiuunerani norraat amerlassusiat appasippoq 14 %-imiilluni, taamatullu aamma kulavaat 100-gaangata norraat amerlassusiat 24-ulluni. Pannerit arnavissallu nikingassusiat tamaallaat 0,45-juvoq. Toquinnartartut amerlassusiat 8-10 %-imiippat tuttoqarfimmi Akia-Maniitsutut ittumi toquinnartartut amerlassusiat ukiumut 2.200 aamma 4.500 akornanniissapput. Tuttoqarfik Kangerlussuaq-Sisimiut – Avannaa Kalaallit Nunaata tuttoqarfittut immikkoortua, tuttoqarfik Kangerlussuaq-Maniitsoq, martsimi 2005 tuttut piaqqiulinnginnerisa nalaani amerlassutsimikkut missingiunneqarput 90.464-inut (70.276-113.614; 90 % KI). Tamaani tuttut amerlanerpaaffiini kvadratkilometerimut amerlassusiat 6,2-miippoq, ikinnerpaaffiinilu 2,3-miilluni. Agguaqatigiissillugu ataatsimoortut 2005-imi amerlassusiat 4,63±3,38 SD-miippoq. 2005-imi ukiuunerani piaqqat amerlassusiat appasippoq 11 %-imiilluni, taamatullu aamma arnavissat 100-gaangata piaqqat amerlassusiat 16-iulluni. Pannerit arnavissallu nikingassusiat tamaallaat 0,33-uvoq. Toquinnartartut amerlassusiat 8-10 %-imiippat tuttoqarfimmi Kangerlussuaq-Sisimiutut ittumi toquinnartartut amerlassusiat ukiumut 5.600 aamma 11.400 akornanniissapput. Piaqqat amerlassusiat toquinnartartut amerlassusiat nallersimalermagu imaarataannaavoq tamaani tuttoqassutsip killissani tikissimassagaa, tassa piaqqat toquinnartartullu amerlaqatigiilermata tuttut amerliartorunnaarsimassallutik. Resume I marts 2005 blev bestandstætheden og flokstrukturen af to rensdyrbestande, Kangerlussuaq-Sisimiut og Akia-Maniitsoq, bestemt ved helikopter-optælling. Metoder og analyser var de samme som i Cuyler et al. (2003). Den anbefalede floktæthed er overskredet med ca. 60.000 rener i Kangerlussuaq-Sisimiut populationen og ca. 17.500 i Akia-Maniitsoq popula- 8 tionen. Både andelen af kalve og rekrutteringen er lav i 2005. Der ses også et fald i antal af bukke i forhold til simle. Den lave produktion og den skæve kønsfordeling kan blive mere udtalt fremover. Jagttrykket menes ikke at være en væsentlig årsag til disse ændringer, da det lave fangstantal ikke kan have haft nogen større effekt på så store populationer. Selv om vi mangler kvantitative data, ser det ud til, at tæthedsafhængige effekter forårsager højere dødelighed blandt kalvene, og måske endda også blandt bukkene i Kangerlussuaq-Sisimiut og Akia-Maniitsoq populationen. Idet tæthederne er tre til seks gange højere end den anbefalede målværdi, der anses for at være bæredygtig, forventer vi, at konkurrencen mellem dyrene om de tilgængelige føderessourcer er hård. Hvis bestandenes størrelse får lov at fortsætte på det nuværende niveau eller øges yderligere, er der en klar risiko for varige skader på græsningsarealerne, fx i form af overgræsning og nedtrampning. Ikke-bæredygtig udnyttelse af græsningsarealerne kan blive en trussel for rensdyrbestandenes fremtidige sundhedstilstand og levedygtighed i Vestgrønland. Uanset hvilke forvaltningsmæssige forholdsregler der tages nu, kan bestandssammenbrud være uundgåeligt for nogle af Vestgrønlands rensdyrflokke inden for en overskuelig fremtid, men det er umuligt at forudsige flokudviklingen præcist. Skal vi forstå den fremtidige udvikling, må rensdyrene og deres græsningsarealer studeres i en bredere sammenhæng, som også omfatter den globale opvarmning og de medfølgende klimaforandringer. Akia-Maniitsoq bestanden – Region Midt Akia-Maniitsoq bestanden i Region Midt anslås i marts 2005 at have en populationsstørrelse før kælvning på ca. 35.807 rener (27.474-44.720; 90 % KI). Rensdyrtætheden var i 2005 på 3,0 rener pr. km2 hvor tætheden var størst, og 1,1 pr. km2 hvor tætheden var mindst. Den gennemsnitlige flokstørrelse var 4,33 ±2,91 SD i 2005. Senvinter-andelen af kalve var lav, 14 %, og ligeså den årlige rekruttering på 24 kalve pr. 100 simle. Forholdet mellem bukke og simle var på kun 0,45. Ved en naturlig dødelighed på 8-10 %, vil man i en flok på denne størrelse kunne forvente at se en naturlig dødelighed på mellem 2200 og 4500 dyr om året. Kangerlussuaq-Sisimiut bestanden – Region Nord Kangerlussuaq-Sisimiut bestanden anslås i marts 2005 at have en populationsstørrelse før kælvning på ca. 90.464 rener (70.276-113.614; 90 % KI). Rensdyrtætheden var i 2005 på 6,2 rener pr. km2 hvor tætheden var størst, og 2,3 pr. km2 hvor tætheden var mindst. Den gennemsnitlige flokstørrelse var 4,63 ±3,38 SD i 2005. Senvinter-andelen af kalve var lav, 11 %, og ligeså den årlige rekruttering på 16 kalve pr. 100 simle. Forholdet mellem bukke og simle var på kun 0,33. Ved en naturlig dødelighed på 8-10 %, vil man i en flok på denne størrelse kunne forvente at se en naturlig dødelighed på mellem 5600 og 11.400 dyr om året. Da andelen af kalve nærmer sig den naturlige dødelighed, kan denne bestand være på vej mod sin teoretiske bærekapacitet, hvor fødsler er lig med dødsfald, dvs. produktionen er lig nul. 9 Introduction Caribou (Rangifer tarandus groenlandicus) have no natural predators in West Greenland, and none have existed for several hundred years (Dawes et al. 1986). When combined with their high fertility (Cuyler & Østegaard 2005) and recruitment (Cuyler et al. 2002, 2003, 2004), this would suggest that overabundance may be their greatest threat. Several boom and crash cycles of caribou in West Greenland have been noted since the 1700’s (Vibe 1967, Meldgaard 1986), and recent population estimates are the highest ever documented, indicating that a new crash might be expected in the near future. Past population estimates Total caribou abundance in West Greenland may have been about 100,000 animals in the late 1960’s with a proposed crash to about 16,000 animals by 1978 (Clausen et al. 1980, Roby & Thing 1985). Following unsystematic surveys the estimates were 79,000 in 1980 and 15,000 in 1982 (Strandgaard et al. 1983). During that period there was little correlation between the population estimates and government harvest statistics. Over 6,000 caribou were harvested in 1980, and over 9,000 in both 1982 and 1983 (Born et al. 1998). Those harvest numbers would not have been possible if the aerial survey estimates of 1980-82 had been close to accurate. No critisicm of the population estimates occurred, perhaps because the public was unaware of them and hunting remained unregulated. Systematic aerial surveys completed in the 1990’s suggested that caribou in West Greenland were few in number. In 1993 the estimate was about 7-9,000 caribou, and in 1996 about 20-22,000 (Ydemann & Pedersen 1999). Local knowledge contradicted the low estimates. Therefore, the accuracy of these estimates was hotely debated and created much public anger because hunting was first prohibited for 2 years and then heavily regulated for the first time. All surveys have intrinsic errors and biases. Given the methods employed in the 1990’s (high speed, high altitude, wide strip width, long transect length, sun glare, inability to maintain constant altitude, etc.), it is likely that these surveys underestimated populations in West Greenland because a large number of caribou present within the area of the transects were not seen (Cuyler et al. 2002, 2003). 16 Area 2 Area 2 Area 2 Area 2 Area 2 Area 2 Area 2 Area 2 Area 2 Area 6 Area 6 Area 6 Area 6 Area 6 Area 6 Area 6 Area 6 Area 6 Area 1 Area 1 Area 1 Area 1 Area 1 Area 1 Area 1 Area 1 Area 1 Area 3 Area 3 Area 3 Area 3 Area 3 Area 3 Area 3 Area 3 Area 3 Area 4 Area 4 Area 4 Area 4 Area 4 Area 4 Area 4 Area 4 Area 4 Area 5 Area 5 Area 5 Area 5 Area 5 Area 5 Area 5 Area 5 Area 5 Eldorado Eldorado Eldorado Eldorado Eldorado Eldorado Eldorado Eldorado Eldorado Maniitsoq Maniitsoq Maniitsoq Maniitsoq Maniitsoq Maniitsoq Maniitsoq Maniitsoq Maniitsoq 56 17 164 39 1 87 200 Nuuk Nuuk Nuuk Nuuk Nuuk Nuuk Nuuk Nuuk Nuuk 124 35 35 35 35 35 35 35 35 35 0 0 0 0 0 0 0 0 070 70 70 70 70 70 70 70 70 kilometres kilometres kilometres kilometres kilometres kilometres kilometres kilometres kilometres 136 Figure 4. Central region: Akia-Maniitsoq herd structure zigzag overflight areas (indicated by blue crosshatching) and transects (the blue transects with ID number highlighted were zigzagged; the red transects indicate where opportunistic observations where obtained). The high caribou density stratum included the area inside the blue outline. Elevation is not shown. 143 K20 K20 K20 K20 K20 K20 K20 K20 K20 29 203 P21 P21 P21 P21 P21 P21 P21 P21 P21 S19 S19 S19 S19 S19 S19 S19 S19 S19 122 115 172 65 209 211 W20 W20 W20 W20 W20 W20 W20 W20 W20 200 137 92 197 112 210 24 116 36 34 58 104 189 154 106 73 9 32 153 192 70 59 192 153 153 8 Sisimiut Sisimiut Sisimiut Sisimiut Sisimiut Sisimiut Sisimiut Sisimiut Sisimiut 0 0 0 0 0 0 0 0 0 kilometres kilometres kilometres kilometres kilometres kilometres kilometres kilometres kilometres 30 30 30 30 30 30 30 30 30 60 60 60 60 60 60 60 60 60 Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Kangerlussuaq Figure 5. North region: Kangerlussuaq-Sisimiut herd structure zigzag overflight areas (indicated by blue crosshatching) and transects (the blue transects with ID number highlighted were zigzagged; the red transects indicate where opportunistic observations where obtained). The high caribou density stratum included the area inside the blue outline. Elevation is not shown. 17 during zigzagging. All caribou sighted were sexed and aged (< or > 1 year old) following a brief overpass with the helicopter. Sex was determined by the presence or absence of a vulva and/or urine patch on the rump. This reliably indicated a female on both adults and calves. No other method was 100% certain, e.g. antler size, shape, presence or absence, were not used, as the presence of antlers on female caribou is highly variable in western Greenland. Age was determined by body size. Calves of both sexes were considerably smaller than all other age classes at this time of year. There were two age classes used in subsequent analyses, i.e. calf (≤ 9-10 months old) and adult (> 1 year). Calf percentage given is the percentage of the total number of caribou seen. Calf recruitment is the late-winter calf per 100 cow ratio. Group size was based on proximity and group cohesion during possible flight response. Results Caribou on the transects, commonly reacted with flight when the helicopter flew by, however, frequently animals remained lying down or standing/grazing (Appendix 12, Figure 92). These animals typically looked at the helicopter but other overt reactions were not forthcoming. Hence movement was not the only key for locating animals present on a transect. The ability to spot the shape or colouring of a stationary caribou was necessary, regardless of the degree of camouflage against the varied backgrounds. The March 2005 snow cover conditions in the Central (Appendix 9a) and specifically the North region (Appendix 9b) made it clear that detecting caribou would have been difficult to impossible if it were not for the survey’s low flight altitude, low speed, and narrow strip width. Current survey design promotes spotting caribou (Appendix 7). No dead caribou were observed. At the flight altitude used, 15 m, “dead” ground is common on transects, i.e. terrain features prevent seeing the entire 300 metre strip width. Caribou may be missed because they are hidden from view. This is a source of negative bias and contributes to under estimating population size. The correction calculation accounted for different correction factors for each stratum. Since no good method is available which could include the variance of a correction factor, the confidence intervals were instead calculated using a bootstrap method (Effron & Tibshirani 1993). 18 Akia-Maniitsoq estimated population size, Central Region We observed a total of 554 caribou. The raw data (Appendix 4) gave an uncorrected pre-calving population estimate of ca 33,181 caribou, with densities of ca 1-3 caribou per sq km for the low and high-density strata respectively. After incorporating a correction for missed caribou (Cuyler et al. 2002), the pre-calving population size estimate for March 2005 became ca 35,807 (90% CI: 27,474 – 44,720), while densities remained basically unchanged (Table 2). The survey of of Akia-Maniitsoq used 19 hours and 57 minutes of flying time. Weather conditions between the 14 and 16 March were excellent for strip visibility and caribou sightability. Snow cover, however, was patchy and could vary dramatically along an individual transect. This increased the difficulty in spotting caribou, and observers had to remain concentrated and focused while counting. Snow cover in the high-density stratum varied between 10 and 99%, while the typical range was 40 to 80%. As usual, the low-density stratum was almost totally covered in deep snow. Total number of Akia-Maniitsoq caribou seen per observer were as follows; 312 Christine Cuyler, 300 Rink Heinrich, 206 Johannes Egede. More caribou were observed on the left side of the helicopter than on the right, 336 and 218 respectively. Kangerlussuaq-Sisimiut estimated population size, North Region We observed a total of 1284 caribou. The raw data (Appendix 5) gave an uncorrected pre-calving population estimate of ca 87,244 caribou, with densities of ca 2-6 caribou per sq km for the low and high-density strata respectively. After incorporating a correction for missed caribou (Cuyler et al. 2002), the pre-calving population size estimate for March 2005 became ca 90,464 (90% CI: 70,276 – 113,613), while densities remained basically unchanged (Table 3). The survey of Kangerlussuaq-Sisimiut used 26 hours and 38 minutes of flying time. This was greater than the number of hours used for the survey in the Central region, owing to the return helicopter ferry between Nuuk and Kangerlussuaq airports and the greater distances to refueling in the North region. Weather conditions on the first and second day of the survey were excellent, however, snow cover was often completely lacking, and typically patchy at best. Typical snow cover was 10-40%, although occassionally up to 99% . On the third and fourth days of the survey, a light dusting of new fall snow produced a “salt & pepper” background, against which the caribou were optimally camouflaged. Snowflurries and low cloud-fog further increased the difficulty in spotting caribou, by obscuring strip width visibility and often creating white-out conditions. These factors increased the 19 difficulty in spotting caribou, and observers had to be extemely concentrated and focused while counting. Total number of Kangerlussuaq-Sisimiut caribou seen per observer were as follows; 701 Hans Mølgaard, 588 Christine Cuyler, 539 Rink Heinrich. More caribou were observed on the left side of the helicopter than on the right, 691 and 593 respectively. Herd structure & recruitment In March 2005, calf recruitment and number of bulls in the population were poor in both populations studied (Table 4, Appendices 4 & 5). Animals were widely spread throughout both regions, with a mean group size at ca 4.6 ±3.4 SD in the Kangerlussuaq-Sisimiut stock, and ca 4.3 ±2.9 SD in the Akia-Maniitsoq stock. Large congregations of animals were not common, and the largest numbered 17 caribou in both regions. Table 2. Survey information and preliminary raw and corrected population size estimates for Akia-Maniitsoq caribou, Central region, 14-16 March 2005. Parameter High-density Low-density Totals Area size 10,037 km2 5,325 km2 15,362 km2 Number strips 39 15 54 Length of each strip 7.5 km 7.5 km Total strip width 2x 300 m 2x 300 m Area covered 175.5 km2 67.5 km2 243 km2 Flight height 15 metres 15 metres Flight speed (km/hr) 46 to 65 46 to 65 Total caribou seen (n) 485 69 554 Raw Density (caribou / km2)* 2.76 1.02 1 to 3 Raw estimate herd size* 27,738 5,443 33,181 Corrected Density (caribou / km2)** 3.00 1.06 1 to 3 Corrected estimate herd size** 30,153 5,654 35,807 90% Confidence Interval (CI) 22,088 – 39,266 3,765 – 7,663 27,474 – 44,720 * Herd size estimate from raw data with no correction for missed caribou. ** Herd size estimate after correction for missed caribou has been made. 20 Table 3. Survey information and preliminary raw and corrected population size estimates for KangerlussuaqSisimiut caribou, North region, 18-22 March 2005. Parameter High-density Stratum Low-density Stratum Totals Area size 8,000 km2 18,000 km2 26,000 km2 Number strips 40 20 60 Length of each strip 7.5 km 7.5 km Total strip width 2x 300 m 2x 300 m Area covered 180 km2 90 km2 270 km2 Flight height 15 metres 15 metres Flight speed (km/hr) 46 to 65 46 to 65 Total caribou seen (n) 1090 194 1284 Raw Density (caribou / km2)* 6.06 2.16 2 to 6 Raw estimate herd size* 48,444 38,800 87,244 Corrected Density (caribou / km2)** 6.22 2.26 +2 to +6 Corrected estimate herd size** 49,723 40,741 90,464 90% Confidence Interval (CI) 41,833 – 58,470 22,263 – 62,251 70,276 – 113,613 * From raw data with no correction for missed caribou. ** After correction for missed caribou has been made. Table 4. Herd Structure for two caribou herds in West Greenland, March 2005. Parameter Akia-Maniitsoq Caribou Population Kangerlussuaq-Sisimiut Caribou Population Region (Hunting area) Central (3) North (2) Time period 14-16 March 2005 18-22 March 2005 Method Helicopter Helicopter Total sexed & aged (n) 705 745 Number of groups observed 163 161 Average group size 4.33 ± 2.91 SD 4.63 ± 3.38 SD Maximum group size 17 17 Minimum group size 1 1 Bull (> 1 year) 187 (26.52%) 163 (21.9%) Cow (> 1 year) 419 (59.43%) 501 (67.3%) Calf from 2004 99 (14.04%) 81 (10.9%) Recruitment (calf/100cow) 24 16.2 Bull to Cow ratio 0.45 0.33 21 Discussion Kangerlussuaq-Sisimiut population The corrected pre-calving March 2005 Kangerlussuaq-Sisimiut population estimate is ca 90,464 caribou (90% CI : 70,276 – 113,614). This is best considered a conservative estimate since a negative bias of caribou missed remains, owing to this year’s weather conditions, patchy snow cover, “salt & pepper” backgrounds, and “dead” ground. This estimate is almost double the 2000 survey estimate and is greater than any previous estimate for this herd. An interpretation of population trend from this result is difficult since methods differed. The present survey better reflects true animal abundance in 2005. The stocking density is now over six caribou per sq km, almost a doubling since the 2000 survey. Given the large population size, it was not unexpected that mean group size increased (Figure 6). Large aggregations of caribou, however, were not common and the maximum group size was unchanged from the survey in 2000. Meanwhile, the percentage and recruitment of calves into the population is the lowest observed (Figure 7, 8). The percentage of calves is similar to the natural mortality value. It is possible that this population is approaching its theoretical carrying capacity where births equal deaths, i.e. production equals zero. A population at carrying capacity will not be able to provide a sustainable optimal yield for annual harvest. Table 5. Greenland caribou population estimates, harvest quotas, reported harvest and the percentage by which the quota was filled. Year Estimate of total caribou in Greenland Quota Reported Harvest (Piniarneq) ** Amount of quota filled 1995 ca 18,000 2,000 1,398 69.9% 1996 ca 22,000 2,600 2,048 78.8% 1997 3,111 2,755 88.6% 1998 3,680 3,692 100.3% 1999 4,050 3,957 97.7% 2000 13,600 9,671 71.1% 2001 ca 140,000 24,300 13,490 55.5% 2002 36,150* 16,910 52.3% 2003 Open 18,851 - 2004 Open Not yet available - * The 2002 harvest quota was originally set at 32,150 caribou; however, the number of licences permitted exceeded that number by 4,000. ** Piniarneq records are from the Directorate for Fisheries & Hunting, P.O. Box 269, 3900 - Nuuk, Greenland. 22 Given that the North region is ca 26,000 sq km, if the recommended stocking density of 1.2 caribou per sq km was attained, then an appropriate population size might be ca 31,200 caribou. The 2000 population estimate for the North region exceeded this by almost 20,500 animals, and the 2005 estimate exceeds it by almost 60,000 caribou. Although annual harvest results per population are unavailable, the total numbers of caribou harvested in Greenland (Table 5) are low relative to the numbers required for population reduction even in just the North region. Therefore, despite the large quotas in 2000-2001 followed by open harvests in 2002/03/04, hunting has not been sufficient to halt growth or reduce the Kangerlussuaq-Sisimiut herd size. Akia-Maniitsoq population The Akia-Maniitsoq stock situation is slightly different. The corrected pre-calving March 2005 Akia-Maniitsoq population estimate is ca 35,807 caribou (90% CI : 27,474 – 44,720). This is also best considered a conservative estimate since a negative bias of caribou missed remains, owing to this year’s patchy snow cover and “dead” ground. This estimate is ca 10,000 animals less than the 2001 survey estimate, but is still a large number for this herd. The only difference between the 2001 and 2005 survey was the addition of seven transects, which could not be flown in 2001 owing to financial constraints. The additional transects in 2005 served only to reduce the variance and did not affect the population size estimate. Since methods between the surveys of 2001 and 2005 did not differ, the present results may reflect a true decrease in animal abundance over the past four years. When the two estimates are compared the result is a P = 0.12. This indicates a 12% probability that the 2005 estimate was smaller by chance. Given that the Central region is ca 15,362 sq km, if the recommended stocking density of 1.2 caribou per sq km was attained, then a suitable population size might be ca 18,434 caribou. The 2001 population estimate for the Central region exceeded this by almost 28,000 animals, and the 2005 estimate exceeds it by almost 17,500 caribou. The Akia-Maniitsoq stock density dropped since 2001, however, it remains too high in 2005. In 2001, it was four-times the recommended target and in 2005 is three-times. Since hunting had no clear impact on the Kangerlussuaq-Sisimiut population, the apparent stock reduction that has occurred in Akia-Maniitsoq is not automatically assumed due to hunting. Rather it could be an expression of greatly increased natural mortality due to overstocking. The latter is supported by the poor calf recruitment observed in March 2005. 23 Density-dependent effects Density-dependent effects are the result of intraspecific competition, i.e. between individuals in the same stock. Direct effects typically increase mortality, while delayed effects affect growth and fecundity. With increasing densities, severe weather events can have additional effects on recruitment, with consequences for population stability (Skogland 1985). In 2001, although caribou densities were as high as 3-4 per sq km, there was no evidence of dramatic density effects in any of the west-coast populations. Still the calf percentage was considered low in two populations, the Akia-Maniitsoq (Central region) and Ameralik (region South). In contrast, the current March 2005 surveys strongly suggest that density dependant factors now play a major role in caribou population dynamics in the two stocks examined. Other stocks in West Greenland may be experiencing similar problems. Recruitment The Akia-Maniitsoq (Central region) had a late winter calf percentage of ca 14% and a recruitment of only 24 calves per 100 cows. The Kangerlussuaq-Sisimiut (North region) stock had a calf percentage of ca 11% and a recruitment of only about 16 calves per 100 cows. The results for both stocks are low compared to herds elsewhere. Studies from North America and Scandinavia report late winter recruitments of 41 calves per 100 cows (Fancy Whitten & Russell 1994), 20 calves per 100 cows (Dzus 1999) and 22 calves per 100 cows (Parker 1972), and some of these populations typically have predators. Further a comparison to the Southampton Island herd, which like Greenland has no predators, shows late winter recruitments varying between 22 and 77 calves per 100 cows (Heard & Ouellet 1994). This suggests that the current Kangerlussuaq-Sisimiut caribou herd late winter recruitment is low, while the Akia-Maniitsoq recruitment is tending in the same direction. The poor calf recruitment strongly suggests an elevated natural mortality among calves, and a decreased fecundity of adult females. Independent of climate and genetics, caribou calf mortality increases with high population density and grazing pressure (Valkenburg et al. 2000). Further, calf recruitment is low or variable where winter ranges are overgrazed and hard or deep snow is common (Heggberget et al. 2002). At Kangerlussuaq snow is ruled out as a cause of decreased recruitment, because hard or deep snow is almost never a problem in the dry steppe climate of the North region, which is where caribou density is highest. Knowledge on possible changes in female fecundity is not available, however, increased calf mortality may occur when animal densities are high. Thing & Clausen (1980) suggested high caribou density increased faeces contamination (bacteria and parasites) of the 24 y = 0.1116x - 219.2 R2 = 0.1303 y = 0.1623x - 321.19 R2 = 0.6649 0 1 2 3 4 5 6 7 1992 1994 1996 1998 2000 2002 2004 2006 Year Mean Group Size Figure 6. Changes in mean group size since 1992 in two West Greenland stocks; Kangerlussuaq-Sisimiut stock (●, - - -) (p = 0.03), and Akia-Maniitsoq stock (□, ____) (p = 0.43); linear regression lines with r2 values. y = -0.2927x + 603.11 R2 = 0.047 y = -0.8254x + 1670 R2 = 0.7218 0 5 10 15 20 25 30 1985 1990 1995 2000 2005 2010 Year Calf percentage Figure 7. Changes in calf percentage since 1990 in two West Greenland stocks, Kangerlussuaq-Sisimiut stock (●, - - -) (p = 0.68), and Akia-Maniitsoq stock (□, ____) (p = 0.03); linear regression lines with r2 values. 25 y = -5.6351x + 11318 R2 = 0.8142 y = -5.6923x + 11434 R2 = 0.6123 0 10 20 30 40 50 60 70 80 1996 1998 2000 2002 2004 2006 Year Calves per 100 Cows Figure 8. Changes in late winter recruitment (calves per 100 cows) in two West Greenland stocks, Kangerlussuaq-Sisimiut stock (●, - - -) (p = 0.43), and Akia-Maniitsoq stock (□, ____) (p = 0.28) ; linear regression lines with r2 values. y = -0.0804x + 161.52 R2 = 0.9477 y = -0.0653x + 131.28 R2 = 0.8921 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1996 1998 2000 2002 2004 2006 Year Bull to Cow Ratio Figure 9. Changes in the Bull to Cow ratio in two West Greenland stocks, Kangerlussuaq-Sisimiut stock (●, - - -) (p = 0.15), and Akia-Maniitsoq stock (□, ____) (p = 0.21) ; linear regression lines with r2 values. 32 Appendix 1 Region Stratification & Transect Allocation How many transects are needed? One of the most important questions that have to be answered before undertaking any survey is whether the survey will yield data of a sufficient quality to answer the question that the survey attempts to answer; animal abundance. A related question is the choice of sample size. In a helicopter survey, where flight hours in Greenland are very expensive, this question becomes very important. An idea of the expected variance is necessary. In flight surveys the variance is intimately related to the density of animals. The prior information available before the surveys was relative densities from a previous survey in 1996 and densities found in the North region in 2000. The assumption made was that although the 1996 surveys used a radically different methodology, the relative densities would remain fairly constant. Implicit in that assumption is the expectation that the caribou populations in all regions have had similar growth rates since 1996 despite that they form clearly distinct populations with different demographics. A simulation experiment was performed in the following fashion. The highest density area in the 1996 survey was the high-density area of the North region, the density of the other areas was known as a fraction of the density of that high-density area. For each simulated transect the number seen is found as follows. A random transect from the high-density area in the North region is chosen and the number seen there is called “s”. If r is the relative density of the area in question and w is the relative width of the transects then a number seen can be simulated as a binomial: (, ) B inomial s r w⋅ Once a simulation was done, the resulting data was analyzed using standard parametric methods, and a confidence interval found. The procedure was repeated for different total numbers of transects. The data was then plotted by taking all the confidence intervals, centering these on their common mean and plotting them against the total number of transects (Figure 10). 33 Figure 10. How many transect lines needed for a relatively accurate and precise survey of the Central region? Simulation of confidence interval mean values versus the number of transects used. From the graph it is obvious that an effort smaller than 40 lines will result in a wide confidence interval, whereas a number larger than 60 will be a waste of resources. Note that the picture here is slightly misleading since it takes into account only the width of the confidence interval around the grand mean of the estimates. In reality the means will jump around less for higher sample sizes. For economic reasons the final number of transect lines in 2005 was set to 54 for the Akia-Maniitsoq herd in the Central region, while 60 transects were again applied for the Kangerlussuaq-Sisimiut herd in the North region. Transect allocation Since the Central region is divided into two strata with different expected densities, transect allocation must be decided. Here a simple mathematical method was used for allocating transects to each strata. The standard method for allocation of transects to strata is to allocate proportional to the product of the area and the expected standard deviation of each strata. If : i Ais the area of strata i i dis the expected density of strata i then the best allocation is proportional to total number of transects 20 30 40 50 60 70 80 12000 14000 16000 18000 20000 22000 24000 Confidence interval mean value Number of transects 34 ii Ad α ⋅ where: 0.5 α = corresponds to the square root of the expected density. Note that it is sufficient to have the expected relative densities and areas. For areas {1,..,i} the proportions of transects allocated to area 1 will be. 11 1 1 11 1 11 11 1 ii ii i i ii iii i Ad pAd Ad Ad Ad A Ad d Ad α αα α α αα ⋅ == = = ⋅⋅ ⎛⎞⎛⎞ ⋅⋅ ⎜⎟⎜⎟ ⋅⎝⎠⎝⎠ ∑∑∑∑ There are several ways of choosing α. For animals that tend to be in groups the question centres around whether they tend to increase the group size when the density is higher. If the group size is the same regardless of density then α = 0.5. If on the other hand the group size tends to go up with higher density without the number of groups changing then α = 1. In this case we chose α = 0.75 as a compromise solution. The allocation assumed that the relative densities remained unchanged since last survey of 1996. The stratification was not the same as in the last survey, but was altered based on the observed densities in 1996. The Central and North regions were divided into two strata, a high and low-density strata. On the basis of the above mentioned formulas, in 2005 the Central region was allocated 15 transects to the lowdensity area and 39 transects to the high-density area. Similarly the North region was allocated 20 transects to the low-density area and 40 transects to the highdensity area. 35 Appendix 2 Survey field method and statistical design Accuracy equates to the population size calculated being close to the true value. Bias, which makes the calculated population size depart from reality, results in inaccuracy. There can be bias in your counting, sampling design or even analysis. Precision is the measure of variation in the numbers of caribou on each of the transects. Poor precision can result from sampling errors, e.g. if group size and distribution were highly variable within a stratum. Field methods – Reducing negative bias: Sightability of caribou on transect To reduce the negative bias associated with violation of the assumption that all caribou within the strip are observed, the following survey field methods were used. Narrow strip width, 300x2 metres, Slow flying speed, ca 46-65 kilometre/hour, Low altitudes, 15 metres, Sun typically behind observers, Short transect length, 7.5 kilometres (promoted concentration and reduced fatigue), Statistical correction for missed caribou. Statistical design Caribou population estimates can be calculated as follows: For each stratum we have: ˆij i jj i i yA NA y AA =⋅ = ⋅ ∑ ∑ (0.1) Where ˆ j N is the estimated total in the jth strata i y is the total number of caribou observed in strip i j A is the total area of strata j i A is the area of strip i A is the mean area of the strips in the stratum Because the area of each strip is constant the calculation of variance is 36 22 2 22 2 22 ˆ () ( ) 1 () ( ) ( ) () 11 () ( ()) j j i jj j ii i jj j i ii i A Var N Var y A y AA A Var y Var Var y AAnAn AA A Var y Var y n Var y An An A n =⋅= ⎛⎞ ⎛⎞ ⎛⎞ =⋅ =⋅ = ⎜⎟ ⎜⎟ ⎜⎟ ⎝⎠ ⎝⎠ ⎝⎠ ⎛⎞ ⎛⎞ ⎛⎞ ⋅=⋅⋅=⋅ ⎜⎟ ⎜⎟ ⎜⎟ ⎝⎠ ⎝⎠ ⎝⎠ ∑∑ ∑ 22 1 ˆ() ( ) 1 ii i Var y s y y n == − −∑ Since the total number of caribou in the area is the sum of the totals in each stratum the variance of the total will be the sum of the variances in the strata. ˆj j j A Ny A =⋅ ∑ 2() ˆ () ji j AVar y Var N An ⎛⎞ =⋅ ⎜⎟ ⎝⎠ ∑ 37 Appendix 3 Increasing the accuracy of aerial counts of caribou in western Greenland. Most aerial surveys of animal abundance are negatively biased because animals within the sample unit are overlooked by observers. Various doublecount methods have been developed to generate survey specific correction factors. However, these methods require that observations can be attributed to specific individuals or groups, which is not always possible. We present a simple method for generating a minimum estimate of the number of overlooked animals based on the total number of animals seen by double observers on one side of the aircraft. In addition, we describe aspects of survey design that have been used in caribou Rangifer tarandus surveys in West Greenland to further reduce bias. The extent to which animals are overlooked can be influenced by many factors such as aircraft design, flying speed, flight height, light conditions, vegetation density, topographic complexity, and observer experience/fatigue (Caughley 1974, Samuel et al. 1987, Aastrup & Mosbech 1993). Early attempts to correct for this bias focused on determining a factor from a series of controlled trials, and using this as a blanket correction factor for all further surveys (Caughley 1974, Caughley et al. 1976, Samuel et al. 1987, Pollock & Kendall 1987, Aastrup & Mosbech 1993). However, because conditions vary from survey to survey there have been attempts to develop surveyspecific correction factors, especially using the doublecount methodology (Pollock & Kendall 1987, Graham & Bell 1989, Rivest et al. 1995). In this process, at least one side of the aircraft has two observers. Using the numbers of animals or groups seen by the first observer only, the second observer only, or by both observers it is possible to apply capture-mark-recapture methodology to calculate the number of animals seen by neither observer (Pollock & Kendall 1987). However, this requires that observations from the two observers can be attributed specifically to each animal or group observed. While such results may be achieved using double-track tape recorders (Marsh & Sinclair 1989) or GPS/data logger technology, there are always situations whereby technology fails, is unavailable or cannot be applied practically. We present an extension of the normal doublecount statistics to estimate the correction factor for the proportion of animals unseen using the total number of animals counted by each observer within a given sample strip. In many ways this is similar to the aims of Caughley & Grice (1982), but is designed for species that occur at a higher density. 38 Accounting for overlooked animals In the cases where there are more than one observer in one side of the aircraft and it is possible to know which animals have been seen or not seen by each observer, it is possible to estimate the probability that a visible animal has been observed. The method is thoroughly discussed in Pollock & Kendall (1987) and will be slightly elaborated upon here. We will use the following nomenclature similar to the one used by Graham & Bell (1989). B is the number of animals observed by both observers f S is the number of animals observed by the front seat observer only r S is the number of animals seen by the rear seat observer only M is the number of animals not seen by either observer f p is the probability that a visible animal is seen by the front seat observer r p is the probability that a visible animal is seen by the rear seat observer N is the total number of visible animals in the transects Then fr NS S BM=+++ In a conventional doublecount setup where animals or groups can be individually identified for comparison between observers the following procedure is often used: B can be estimated as () fr EB p p N=⋅⋅ Therefore () f r EB N p p = ⋅ In the same manner f S can be estimated as () (1 ) ff r ES p p N=⋅− ⋅ By substitution () () (1 ) () ()(1 ) () () () (( ) ()) () () () ( ) ff r f r fr r f rr fr r f EB ES p p p p EB ES p p ES p EB EB p ES EB p EB EB pEB ES =⋅− ⋅ ⋅ =− ⋅ ⋅= − ⋅ +⋅= = + In the same manner f p can be estimated as 39 () () ( ) f r EB pEB ES = + Thereby the proportion of animals overlooked by both the front and the rear seat observer is (1 ) (1 ) f r p p−⋅− Therefore, the number of observed animals in the left side of the helicopter should be multiplied with ()() 11 1(1 )(1 ) ( ) 1(1 )(1 ) f r fr fr rf B SBS BB p pBBSS BS BS +⋅+ == −− ⋅− ⋅ + + −− ⋅− ++ or equivalently ()()()() ˆ() () f rfr fr fr B SBS BSBS NBSS BBS S B +⋅+ +⋅+ =++ ⋅ = ⋅+ + And, under the assumption that the left and right rear seat observers have the same probability of observing a visible animal, the right side observations should be multiplied by 1 f r B S p B + = This method does not take into account the variance in the estimates of f p and r p . The easiest way to find confidence intervals is to use a bootstrap procedure (Effron & Tibshirani 1993). The estimates of f p and r p are equivalent to the Petersen estimate. Although this estimate is biased, the bias can be eliminated using Chapman’s correction. (1)(1) ˆ1 1 fr left BS BS NB ++⋅++ =− + (Graham & Bell 1989) Then ˆ r N SB+will be an estimate of 1 r p Hence the estimate of the number of animals on the right side of the aircraft is 40 (1)(1)(1) ˆ(1)( ) fr right right r BS BS B NS BSB ++⋅++−+ =⋅ +⋅ + However, if we don’t know which specific animals or groups have been seen by each observer but have the total number of animals observed within each strip for each observer, then we can calculate maximum values for and f r p p If for each strip i i f is the number of animals seen by the observer in the front seat i r is the number of animals seen by the rear seat observer Then we can define * * * (,) (0, ) (0, ) ii i f ii i rii i BMinfr SMaxfr SMaxrf = =− =− ∑ ∑ ∑ and observe that * * * (1 ) (1 ) fr ff r rr f BppN Sp pN Sp pN ≥⋅⋅ ≤⋅− ⋅ ≤⋅− ⋅ leading to * * * * * * and (1 ) (1 ) fr ff r f r f r r f B Npp B Sp pNp p p p B pBS ≤ ⋅ ≤⋅−⋅≤⋅−⋅ ⋅ ≤ + Similarly 41 * * * f r B p B S ≤ + Since we here are dealing with maximum values of and f r p p the corresponding values for overlooked animals will be minimum values. Accordingly the corrected values for the number of animals seen will still be negatively biased. As this methodology gives a lower bound of the probability of observing a visible animal it is instructive to simulate some observations in order to gauge the effectiveness of the method. Since we are assuming that for each transect line the number seen by both observers is equal to the lowest number seen, it would be reasonable to assume that the method works best for small observation numbers and large observation probabilities. This assumption can be tested using a simulation study. In this simulation a number of virtual surveys were set up, each with 100 transect strips. For each assumed level of detection probability (0.6; 0.7; 0.8; 0.9) a mean number of animals per strip was chosen between 1 and 10. The number of animals on each transect strip was chosen as a Poisson random variable. The number of animals seen by each observer was then chosen as a binomial random variable. The resulting estimates of the sighting probabilities were then plotted against the mean number of animals per strip. As expected (Figure A1) the estimated detection probabilities tended to be too high, particularly when the number of animals per strip is high. Reducing bias through survey design The overriding concern with the survey design has been to minimise the number of overlooked animals by flying closer to the ground and concentrating the effort in a narrow strip close to the aircraft. In addition, observer fatigue was minimised by flying many short transect strips, rather than fewer longer strips. It is possible to evaluate the effectiveness of the different experimental protocols by comparing f p and r p between years. In addition, it is instructive to see how large a difference accounting for overlooked animals makes in each case (Table 7). In the 2000 survey (with the higher flight altitude and wider strip) for the Kangerlussuaq-Sisimiut region there was still a large bias that needed to be corrected. In contrast, the 2001 surveys (lower altitude, narrower strip) in the other three regions resulted in a much smaller bias (Table 7). 48 15.03.05 Area 4 - S mouth of Eldorado Valley 2 0 1 1 15.03.05 Area 4 - S mouth of Eldorado Valley 1 0 1 0 15.03.05 Area 4 - S mouth of Eldorado Valley 3 3 0 0 15.03.05 Area 4 - S mouth of Eldorado Valley 2 0 2 0 15.03.05 Area 4 - S mouth of Eldorado Valley 1 0 1 0 15.03.05 Area 4 - S mouth of Eldorado Valley 6 1 3 2 15.03.05 Area 4 - S mouth of Eldorado Valley 3 0 3 0 15.03.05 Area 4 - S mouth of Eldorado Valley 5 1 3 1 15.03.05 Area 4 - S mouth of Eldorado Valley 4 2 2 0 15.03.05 Area 4 - S mouth of Eldorado Valley 4 0 2 2 15.03.05 Area 4 - S mouth of Eldorado Valley 5 1 3 1 15.03.05 Area 4 - S mouth of Eldorado Valley 3 0 3 0 15.03.05 Area 4 - S mouth of Eldorado Valley 3 0 3 0 15.03.05 Area 5 - Small Valley N of muskox pt 11 0 7 4 15.03.05 Area 5 - Small Valley N of muskox pt 2 0 1 1 15.03.05 Area 5 - Small Valley N of muskox pt 2 0 2 0 15.03.05 Area 5 - Small Valley N of muskox pt 1 0 1 0 15.03.05 Area 5 - Small Valley N of muskox pt 7 0 5 2 15.03.05 Area 5 - Small Valley N of muskox pt 5 1 3 1 15.03.05 Area 5 - Small Valley N of muskox pt 6 0 5 1 15.03.05 Area 5 - Small Valley N of muskox pt 9 2 5 2 15.03.05 136 2 0 2 0 15.03.05 Area 4 – near 46 1 1 0 0 15.03.05 Area 4 – near 46 2 0 1 1 15.03.05 Area 4 – near 65 2 0 2 0 16.03.05 Area 6 - between 61-15 2 0 1 1 16.03.05 Area 6 - between 61-15 4 0 4 0 16.03.05 Area 6 - between 61-15 3 2 1 0 16.03.05 Area 6 - between 61-15 4 2 0 2 16.03.05 Area 6 - between 61-15 4 0 0 4 16.03.05 Area 6 - between 61-15 4 0 1 3 16.03.05 Area 6 - between 61-15 3 0 0 3 16.03.05 Area 6 - between 61-15 3 1 1 1 16.03.05 Area 6 - between 61-15 2 1 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 1 1 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 1 1 16.03.05 Area 4 - Highway valley NE of Eldorado 1 0 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 3 0 16.03.05 Area 4 - Highway valley NE of Eldorado 5 1 3 1 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 1 2 16.03.05 Area 4 - Highway valley NE of Eldorado 1 0 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 4 0 4 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 2 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 2 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 2 0 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 3 0 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 3 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 2 0 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 2 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 1 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 1 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 6 0 5 1 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 3 0 16.03.05 Area 4 - Highway valley NE of Eldorado 7 0 7 0 16.03.05 Area 4 - Highway valley NE of Eldorado 6 2 4 0 16.03.05 Area 4 - Highway valley NE of Eldorado 1 0 1 0 16.03.05 Area 4 - Highway valley NE of Eldorado 3 0 3 0 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 1 1 16.03.05 Area 4 - Highway valley NE of Eldorado 2 0 0 2 16.03.05 Area 4 - Eldorado N to middle 2 0 1 1 16.03.05 Area 4 - Eldorado N to middle 1 1 0 0 16.03.05 Area 4 - Eldorado N to middle 8 8 0 0 16.03.05 Area 4 - Eldorado N to middle 4 0 4 0 16.03.05 Area 4 - Eldorado N to middle 3 0 3 0 16.03.05 Area 4 - Eldorado N to middle 2 0 1 1 16.03.05 Area 4 - Eldorado N to middle 4 1 1 2 16.03.05 Area 4 - Eldorado N to middle 5 0 5 0 16.03.05 Area 4 - Eldorado N to middle 6 1 5 0 16.03.05 Area 4 - Eldorado N to middle 5 0 5 0 16.03.05 Area 4 - Eldorado N to middle 7 1 6 0 16.03.05 Area 4 - Eldorado N to middle 7 3 3 1 16.03.05 Area 4 - Eldorado N to middle 5 0 4 1 49 16.03.05 Area 4 - Eldorado N to middle 3 0 3 0 16.03.05 Area 4 - Eldorado N to middle 7 1 5 1 16.03.05 Area 4 - Eldorado N to middle 14 4 8 2 16.03.05 Area 4 - Eldorado N to middle 3 1 2 0 16.03.05 Area 4 - Eldorado N to middle 4 0 4 0 16.03.05 Area 4 - Eldorado N to middle 4 1 3 0 16.03.05 Area 4 - Eldorado N to middle 5 0 5 0 16.03.05 Area 4 - Eldorado N to middle 2 0 2 0 16.03.05 Area 4 - Eldorado N to middle 3 2 1 0 16.03.05 Area 4 - Eldorado N to middle 2 0 2 0 16.03.05 Area 4 - Eldorado N to middle 2 0 2 0 16.03.05 Area 4 - Eldorado N to middle 3 0 3 0 16.03.05 Area 4 - Eldorado N to middle 6 2 4 0 16.03.05 Area 4 - Eldorado N to middle 6 2 4 0 16.03.05 Area 4 - Eldorado N to middle 3 0 3 0 16.03.05 Area 4 - Eldorado N to middle 8 5 3 0 16.03.05 Area 4 - Eldorado N to middle 9 4 5 0 16.03.05 Area 4 - Eldorado N to middle 3 1 2 0 16.03.05 Area 4 - Eldorado N to middle 3 0 3 0 16.03.05 Area 4 - Eldorado N to middle 4 4 0 0 16.03.05 Area 4 - Eldorado N to middle 2 2 0 0 16.03.05 Area 4 - Eldorado N to middle 6 1 5 0 16.03.05 Area 4 - Eldorado N to middle 13 11 2 0 TOTALS 705 187 419 99 50 Appendix 5 Aerial survey 2005 Kangerlussuaq-Sisimiut caribou, North region, West Greenland Table 11. Raw data aerial survey Kangerlussuaq-Sisimiut caribou herd, North region, March 2005. Kangerlussuaq-Sisimiut observers: (CC) Christine Cuyler, (RH) Rink Heinrich , and(HM) Hans Mølgaard. Number Caribou observed on transect Rear Seat Observers Date ddmmyy Transect number Density Stratum Left front (CC) Left rear Right rear Left Right 18.03.05 77 Low 6 6 6 HM RH 18.03.05 64 Low 6 8 2 HM RH 18.03.05 27 Low 0 1 0 HM RH 18.03.05 151 Low 1 0 0 HM RH 18.03.05 161 Low 10 7 2 HM RH 18.03.05 113 Low 1 1 0 HM RH 18.03.05 101 Low 12 14 0 HM RH 18.03.05 47 Low 0 0 1 RH HM 18.03.05 155 Low 0 0 26 RH HM 18.03.05 87 Low 0 0 0 RH HM 18.03.05 29 Low 3 3 0 RH HM 18.03.05 120 High 22 23 19 RH HM 18.03.05 193 High 6 7 9 RH HM 18.03.05 203 High 37 24 37 RH HM 18.03.05 143 Low 0 0 7 RH HM 18.03.05 139 High 13 13 13 RH HM 18.03.05 122 High 4 3 4 RH HM 18.03.05 202 High 3 3 13 RH HM 18.03.05 115 High 12 12 8 RH HM 18.03.05 76 High 9 12 9 RH HM 18.03.05 172 High 36 36 11 RH HM 18.03.05 211 High 6 6 2 RH HM 18.03.05 200 High 13 18 18 RH HM 19.03.05 125 Low 0 0 0 HM RH 19.03.05 32 Low 2 2 6 HM RH 19.03.05 8 Low 28 37 15 HM RH 19.03.05 61 Low 0 1 0 HM RH 19.03.05 135 Low 1 1 4 HM RH 19.03.05 5 Low 14 13 18 HM RH 19.03.05 150 Low 0 1 6 HM RH 19.03.05 158 Low 0 1 0 HM RH 19.03.05 175 High 3 3 4 RH HM 19.03.05 24 High 15 15 10 RH HM 19.03.05 10 High 17 20 23 RH HM 19.03.05 34 High 15 15 19 RH HM 19.03.05 183 High 4 4 3 RH HM 19.03.05 137 High 8 13 39 RH HM 19.03.05 36 High 15 14 9 RH HM 19.03.05 116 High 59 57 23 RH HM 19.03.05 152 High 5 5 13 RH HM 21.03.05 73 High 17 16 7 HM RH 21.03.05 9 High 7 20 22 HM RH 21.03.05 153 High 0 2 6 HM RH 21.03.05 142 High 11 11 12 HM RH 21.03.05 192 High 15 32 12 HM RH 21.03.05 106 High 15 30 14 HM RH 21.03.05 58 High 10 11 7 HM RH 21.03.05 149 High 0 9 9 HM RH 21.03.05 197 High 13 7 11 HM RH 21.03.05 189 High 8 17 0 HM RH 22.03.05 59 High 6 4 6 RH HM 22.03.05 70 High 11 11 17 RH HM 22.03.05 63 High 8 2 6 RH HM 22.03.05 154 High 8 10 23 RH HM 22.03.05 104 High 6 6 8 RH HM 22.03.05 210 High 10 7 7 RH HM 22.03.05 209 High 4 4 3 RH HM 22.03.05 65 High 10 10 24 RH HM 22.03.05 112 High 24 22 5 HM RH 22.03.05 92 High 19 17 15 HM RH TOTAL 1284 (691 left side + 593 right side) 51 Table 12. Random transects for aerial survey Kangerlussuaq-Sisimiut caribou herd, North region, March 2005. Transect start DD mm.m Transect end DD mm.m Date ddmmyy Direction flown Transect number Latitude Longitude Latitude Longitude 19.03.05 NE-SW 5 66º 30.3' 50º 24.5' 66º 28.9' 50º 34.0' 19.03.05 SE-NW 8 66º 44.4' 50º 20.4' 66º 46.5' 50º 29.2' 21.03.05 SW-NE 9 67º 13.0' 50º 20.4' 67º 16.4' 50º 14.4' 19.03.05 SE-NW 10 66º 59.1' 51º 12.8' 67º 01.9' 51º 20.3' 19.03.05 WSW-ENE 24 66º 56.9' 51º 19.6' 66º 55.0' 51º 28.7' 18.03.05 NE-SW 27 66º 35.0' 52º 14.4' 66º 33.1' 52º 23.3' 18.03.05 WSW-ENE 29 67º 26.8' 52º 59.0' 67º 28.5' 52º 49.3' 19.03.05 SE-NW 32 66º 51.7' 49º 56.0' 66º 55.1' 50º 01.7' 19.03.05 SW-NE 34 67º 02.3' 51º 17.5' 67º 06.2' 51º 14.3' 19.03.05 S-N 36 67º 14.2' 51º 05.3' 67º 10.5' 51º 01.3' 18.03.05 ESE-WNW 47 67º 01.5' 53º 31.1' 67º 02.4' 53º 41.2' 21.03.05 W-E 58 67º 15.3' 50º 51.3' 67º 16.9' 50º 41.8' 22.03.05 SW-NE 59 67º 17.2' 49º 59.5' 67º 20.5' 49º 53.3' 19.03.05 SE-NW 61 66º 39.4' 50º 37.2' 66º 42.1' 50º 44.8' 22.03.05 SE-NW 63 67º 39.3' 50º 11.7' 67º 41.5' 50º 20.7' 18.03.05 NW-SE 64 66º 36.6' 51º 53.9' 66º 36.6' 51º 43.7' 22.03.05 NE-SW 65 67º 33.1' 51º 36.5' 67º 36.5' 51º 30.4' 22.03.05 SE-NW 70 67º 21.9' 50º 09.6' 67º 24.9' 50º 16.6' 21.03.05 WNW-ESE 73 67º 09.9' 50º 25.6' 67º 09.2' 50º 15.4' 18.03.05 W-E 76 * 67º 30.7' 51º 52.5' 67º 34.1' 51º 47.2' 18.03.05 SW-NE 77 66º 47.1' 51º 52.8' 66º 43.7' 51º 58.3' 18.03.05 S-N 87 67º 18.8' 53º 01.6' 67º 22.6' 53º 05.2' 22.03.05 NE-SW 92 67º 29.2' 51º 14.7' 67º 26.8' 51º 23.2' 18.03.05 SSE-NNW 101 66º 21.0' 53º 26.8' 66º 24.6' 53º 31.4' 22.03.05 SE-NW 104 67º 40.9' 50º 45.4' 67º 42.6' 50º 55.0' 21.03.05 NE-SW 106 67º 22.4' 50º 42.1' 67º 18.6' 50º 46.3' 22.03.05 S-N 112 67º 27.4' 51º 06.0' 67º 31.4' 51º 08.2' 18.03.05 SE-NW 113 66º 33.7' 53º 08.1' 66º 36.6' 53º 15.1' 18.03.05 NW-SE 115 67º 35.1' 51º 47.6' 67º 38.0' 51º 55.0' 19.03.05 SW-NE 116 67º 04.4' 50º 44.9' 67º 08.2' 50º 41.7' 18.03.05 SSE-NNW 120 67º 27.1' 52º 33.1' 67º 30.9' 52º 36.6' 18.03.05 SSE-NNW 122 67º 26.2' 52º 09.4' 67º 22.8' 52º 03.8' 19.03.05 E-W 125 66º 56.3' 50º 35.8' 66º 55.9' 50º 25.5' 19.03.05 S-N 135 66º 34.5' 50º 32.9' 66º 38.5' 50º 34.3' 19.03.05 SW-NE 137 67º 13.2' 51º 15.0' 67º 17.1' 51º 12.5' 18.03.05 SSE-NNW 139 67º 23.2' 52º 20.7' 67º 19.9' 52º 14.9' 21.03.05 NW-SE 142 67º 18.6' 50º 10.2' 67º 21.5' 50º 17.4' 21.03.05 SSE-NNW 143 67º 20.4' 52º 38.9' 67º 17.0' 52º 33.2' 21.03.05 E-W 149 67º 17.7' 51º 02.7' 67º 17.7' 50º 52.2' 19.03.05 E-W 150 66º 30.9' 50º 49.8' 66º 31.0' 50º 59.9' 18.03.05 SSW-NNE 151 66º 38.6' 52º 34.9' 66º 42.5' 52º 31.9' 19.03.05 WSW-ENE 152 ** 67º 04.3' 50º 41.5' 67º 06.2' 50º 32.3' 21.03.05 SW-NE 153 67º 13.2' 50º 33.2' 67º 16.8' 50º 28.2' 22.03.05 S-N 154 67º 41.9' 50º 32.8' 67º 45.8' 50º 35.2' 18.03.05 WSW-ENE 155 67º 22.9' 53º 37.2' 67º 24.8' 53º 27.9' 19.03.05 SW-NE 158 66º 34.4' 51º 26.5' 66º 37.8' 51º 21.0' 18.03.05 SSW-NNE 161 66º 36.5' 52º 56.1' 66º 40.3' 52º 52.3' 18.03.05 NW-SE 172 67º 30.5' 51º 43.6' 67º 28.7' 51º 34.1' 19.03.05 WSW-ENE 175 66º 48.3' 51º 44.0' 66º 49.6' 51º 34.3' 19.03.05 W-E 183 *** 67º 01.1' 51º 08.6' 67º 02.2' 50º 58.6' 21.03.05 NW-SE 189 67º 25.4' 50º 49.5' 67º 27.2' 50º 58.9' 21.03.05 SE-NW 192 67º 21.4' 50º 30.8' 67º 25.5' 50º 34.3' 18.03.05 SW-NE 193 67º 32.3' 52º 27.0' 67º 34.7' 52º 18.5' 21.03.05 SE-NW 197 67º 21.8' 50º 51.6' 67º 24.3' 50º 59.8' 18.03.05 NW-SE 200 67º 04.9' 51º 22.7' 67º 06.4' 51º 32.4' 18.03.05 SE-NW 202 67º 35.5' 52º 00.8' 67º 32.4' 51º 54.0' 18.03.05 SW-NE 203 67º 31.2' 52º 22.1' 67º 27.3' 52º 24.4' 22.03.05 SW-NE 209 67º 34.6' 51º 22.8' 67º 37.9' 51º 16.4' 22.03.05 NNE-SSW 210 67º 34.3' 51º 11.7' 67º 38.3' 51º 10.7' 18.03.05 NW-SE 211 67º 11.6' 51º 30.9' 67º 14.2' 51º 38.9' * Transect 76 – the original 2000 end point, 67º 32.4'N / 51º 49.9'W made transect too short therefore corrected in 2005. ** Transect 152 – Pilot error in 2005 on key-in of transect end points, caused this end point to actually be 66º 37.1’N; 50º 33.5’W. *** Transect 183 – Typo error in data sheet from 2000 made transect too short. Was 50º 08.6' in table, here corrected to 51º 08.6'. 52 Table 13. Raw data aerial survey herd structure Kangerlussuaq-Sisimiut caribou herd, North region, March 2005. Date ddmmyy Transect number / Area Flown Group Size Males (Age > 1 year) Females (Age > 1 year) Calves (Age < 1 year) 18.03.05 29 3 0 3 0 18.03.05 203 2 0 1 1 18.03.05 203 2 0 1 1 18.03.05 203 3 1 2 0 18.03.05 203 6 2 2 2 18.03.05 203 2 0 1 1 18.03.05 203 6 3 3 0 18.03.05 143 2 0 0 2 18.03.05 139 2 0 1 1 18.03.05 122 1 0 0 1 18.03.05 122 3 0 0 3 18.03.05 203 5 5 0 0 18.03.05 203 3 3 0 0 18.03.05 203 2 0 1 1 18.03.05 203 4 1 2 1 18.03.05 203 4 1 3 0 18.03.05 203 2 2 0 0 18.03.05 203 8 4 4 0 18.03.05 203 3 0 1 2 18.03.05 203 6 0 3 3 18.03.05 203 4 1 3 0 18.03.05 203 13 1 8 4 18.03.05 203 12 2 8 2 18.03.05 203 4 0 2 2 18.03.05 203 5 0 5 0 18.03.05 203 6 0 4 2 18.03.05 202 2 0 2 0 18.03.05 172 1 1 0 0 18.03.05 172 2 0 1 1 18.03.05 172 3 0 3 0 18.03.05 172 9 1 7 1 18.03.05 172 2 0 2 0 18.03.05 172 6 0 6 0 18.03.05 172 17 6 10 1 18.03.05 172 3 0 2 1 18.03.05 211 2 0 2 0 18.03.05 200 2 0 2 0 18.03.05 200 4 0 2 2 19.03.05 32 1 1 0 0 19.03.05 32 1 1 0 0 19.03.05 8 3 0 3 0 19.03.05 8 2 0 2 0 19.03.05 8 2 0 2 0 19.03.05 8 4 2 2 0 19.03.05 8 1 0 1 0 19.03.05 8 2 0 2 0 19.03.05 8 1 1 0 0 19.03.05 8 8 2 6 0 19.03.05 8 8 1 7 0 19.03.05 8 3 0 3 0 19.03.05 8 3 0 3 0 19.03.05 8 4 0 4 0 19.03.05 24 2 0 1 1 19.03.05 24 2 2 0 0 19.03.05 24 4 0 2 2 19.03.05 34 4 1 3 0 19.03.05 34 4 0 3 1 19.03.05 34 5 1 4 0 19.03.05 34 2 0 1 1 19.03.05 34 4 2 1 1 53 19.03.05 137 17 3 11 3 19.03.05 137 11 6 5 0 19.03.05 137 5 3 2 0 19.03.05 137 7 3 4 0 19.03.05 137 2 0 1 1 19.03.05 137 7 6 1 0 19.03.05 137 5 2 1 2 19.03.05 137 5 3 2 0 19.03.05 137 4 2 2 0 19.03.05 137 4 1 3 0 19.03.05 137 1 0 1 0 19.03.05 137 2 0 2 0 19.03.05 137 4 1 3 0 19.03.05 137 11 2 7 2 19.03.05 137 4 2 2 0 19.03.05 36 2 0 2 0 19.03.05 36 3 1 2 0 19.03.05 116 3 0 1 2 19.03.05 116 5 0 5 0 19.03.05 116 8 1 7 0 19.03.05 116 9 0 9 0 19.03.05 116 14 1 13 0 19.03.05 116 14 0 14 0 19.03.05 116 7 0 7 0 19.03.05 116 4 1 3 0 19.03.05 116 4 0 4 0 19.03.05 116 9 3 6 0 21.03.05 73 3 0 3 0 21.03.05 73 5 1 4 0 21.03.05 9 4 1 3 0 21.03.05 192 6 0 5 1 21.03.05 106 6 0 6 0 21.03.05 58 1 1 0 0 21.03.05 197 3 0 3 0 21.03.05 197 2 1 1 0 21.03.05 189 1 0 0 1 21.03.05 153 8 1 7 0 21.03.05 115 4 1 3 0 21.03.05 115 6 0 6 0 21.03.05 115 16 0 16 0 21.03.05 115 3 0 3 0 21.03.05 115 3 0 3 0 21.03.05 115 12 2 10 0 21.03.05 115 4 0 4 0 21.03.05 115 2 0 1 1 21.03.05 115 13 2 11 0 21.03.05 115 13 3 10 0 21.03.05 115 5 2 1 2 22.03.05 59 3 0 3 0 22.03.05 70 4 0 4 0 22.03.05 70 3 0 3 0 22.03.05 154 4 0 4 0 22.03.05 154 2 0 1 1 22.03.05 154 1 0 0 1 22.03.05 154 2 0 2 0 22.03.05 104 2 0 0 2 22.03.05 210 5 5 0 0 22.03.05 209 2 2 0 0 22.03.05 209 2 0 2 0 22.03.05 209 3 1 2 0 22.03.05 65 6 0 4 2 22.03.05 112 5 0 5 0 22.03.05 112 2 2 0 0 22.03.05 92 5 0 4 1 54 22.03.05 92 16 1 15 0 22.03.05 92 4 0 4 0 22.03.05 92 3 0 0 3 22.03.05 92 5 0 5 0 22.03.05 92 2 0 2 0 22.03.05 92 2 0 2 0 22.03.05 92 7 0 7 0 22.03.05 92 3 0 3 0 22.03.05 92 2 0 0 2 22.03.05 112 6 1 5 0 22.03.05 112 5 1 4 0 22.03.05 112 5 5 0 0 22.03.05 112 5 1 4 0 22.03.05 112 7 1 6 0 22.03.05 112 2 0 1 1 22.03.05 112 5 0 4 1 22.03.05 112 2 0 2 0 22.03.05 112 7 1 4 2 22.03.05 112 6 0 6 0 22.03.05 112 8 4 4 0 22.03.05 112 6 0 5 1 22.03.05 112 3 1 2 0 22.03.05 112 6 2 4 0 22.03.05 112 2 2 0 0 22.03.05 Between transects 92 and 137+197 7 2 5 0 22.03.05 Between transects 92 and 137+197 4 4 0 0 22.03.05 S of Isortoq River; WNW of line 116 2 0 1 1 22.03.05 S of Isortoq River; WNW of line 116 5 2 3 0 22.03.05 S of Isortoq River; WNW of line 116 3 2 1 0 22.03.05 W20 (Grid Cell) 7 0 1 6 22.03.05 W20 (Grid Cell) 2 2 0 0 22.03.05 S19 (Grid Cell) 4 3 0 1 22.03.05 P21 (Grid Cell) 1 1 0 0 22.03.05 P21 (Grid Cell) 2 2 0 0 22.03.05 P21 (Grid Cell) 3 3 0 0 22.03.05 P21 (Grid Cell) 2 2 0 0 22.03.05 K20 (Grid Cell) 6 6 0 0 TOTALS 745 163 501 81 55 Appendix 6 Recommendations for future Aerial survey methods & design To ensure that caribou can be spotted, the methods described in this report should be used in future aerial surveys. Further, if financially possible more transects are recommended in the low-density strata, as these would reduce variance. Sighting caribou Although seldom significant (P < 0.05), fewer caribou were always observed on the right side of the helicopter, where only one observer was present relative to the left side of the helicopter, where two observers independently counted animals (Table 14). We suggest that a subconscience element of competition existed between the two left side observers, since their results will be compared against each other. This sharpened their concentration and more caribou present on the transect were spotted. Competition, real or imaginary, may be a method to further reduce the number of missed caribou on a survey. Since we were interested in which observer saw more caribou and not how much more they saw, binomial non-parametric tests were used to test for this possible “competition” effect between observers. We tested the accumulative sum of caribou sighted on the left and right side of the helicopter using all data from the 2001 and 2005 surveys. First we tested the two observers on the left side against each other, secondly the left and right rear seat observers were compared. Between left side observers the difference in spotted caribou was not significant (P = 0.84). However, the difference between left and right rear seat observers was significant (P = 0.02). This result supports our hypothesis that more caribou are spotted when two observers simultaneaously, yet independently, scan the transect strip for animals. Therefore we recommend two observers on both sides of the helicopter for future surveys. This is difficult but not impossible with the AS350 helicopter. The second observer for the right rear side could take the middle seat between transects and move to a cramped crouch for actual transects. Flying time on a transect is six to nine minutes, which is not too long a period to maintain an awkward position. If observers took turns being second observer on the right rear side, than no one person suffered for long periods. Alternatively a larger helicopter with room for three observers left side and two on the right might be used. Given that a correction factor is already applied to the results obtained by three observers, it may be argued 56 that a fourth observer or bigger helicopter may not the improve the accuracy of the final estimate enough to justify the increased expense. Area (km2) calculation Areas given in this report are “flat”, and do not reflect the topographical complicity of the regions or the random transects flown. If a GIS elevation model could be created for the regions / transects, then this would increase the accuracy of the estimated caribou densities. Table 14. Summary of caribou observed on the left and right side of the helicopter. Caribou population surveyed Number caribou observed Left Side Number caribou observed Right Side Significance4 P(T ≤ t) 2-tail March 2000 Survey1&2 Kangerlussuaq-Sisimiut 619 386 P = 0.001 Neria 201 131 P = 0.083 March 2001 Survey3 Akia-Maniitsoq 335 296 P = 0.576 Ameralik 343 289 P = 0.417 Qeqertarsuatsiaat 68 28 P = 0.103 March 2005 Survey4 Akia-Maniitsoq 336 218 P = 0.101 Kangerlussuaq-Sisimiut 691 593 P = 0.223 TOTAL 2593 1941 P = 0.011 1 Cuyler et al 2002; 2 Cuyler et al 2004; 3 Cuyler et al 2003; 4 t-Test paired two sample for means. Logistics Tips • Refueling is not always possible between 09:00 and 17:00, Monday to Friday, specifically at Sisimiut airport, which can close early, e.g. 14:00, and possibly also at Maniitsoq. Telephone on the specific day to obtain update on whether refueling is possible and when. • Refueling in Kapisillit or Qeqertarsuatsiaat (Fiskenæsset) is only possible if fuel barrels are already there, and pilot has pumping gear onboard. Refueling may take up to two hours if conditions are adverse. • All airports are closed for Sundays and holidays, unless your project is willing to pay to keep them open. • Helicopter pilots are prohibited from flying more than 7 hours per day. Safety considerations would suggest that less than 7 hours is better when flying the low slow transects used in the caribou surveys. 57 • Bring totally non-scratch cloths, which are approved by AirGreenland Helicopter Charter department to wipe condensation off the inside of the helicopter’s front window. • Book the time period for helicopter use well in advance (minimum two months) and check as to whether AirGreenland has other plans for their helicopter or pilot during the time period for the intended survey. One year, AirGreenland neglected to inform us that their pilot was obliged to participate in an AirGreenland pilots training course. This interrupted the survey when weather was optimal. • Make sure the helicopter has a SATELLITE TELEPHONE. For safety reasons helicopter pilots must call-in by radio to AirGreenland every half-hour and give their position. Since radio contact is impossible at the 15 m flight altitudes used during the survey. The pilot must drop what he’s doing and gain altitude until contact is made. This causes delays and can result in wasted time, i.e. extra expense, for the surveys. With a satellite telephone the pilot can make contact with AirGreenland regardless of where we are in the terrain. • Start and end GPS points keyed-in by the pilot should always be checked prior to takeoff. • Check from helicopter GPS that all transects entered have length 7.5 km. • Check that all transects are actually in helicopter GPS. The number of data points may exceed memory of helicopter GPS, which caused all the first transects entered to be erased in 2005. • Check pilot’s print-out of transect points with your own, and pick out discrepancies prior to take-off. • Always carry your original print-out of transect start and end points with you in helicopter for consultation in case the above still does not catch all human errors. 64 Rink Heinrich & Johannes Egede - NUUK 04/04-2005 (in Greenlandic) Tuttunik kisitsineq. Akia – Akia Nord. Ulloq 14/3-2005 16/3-2005 ilanngullugu tuttunik kisitsisoqarpoq. NAPP-mit ukuulluta peqataavugut. Rink Heinrich Johannes Egede (uti) avannamut Nuummiit Maniitsoq tikillugu. Tassani maluginiarparput Arnaviarpassuaqartoq Angutivissallu ikinnerungaatsiaqisut. Siornatigut tusartakkavut umimmaat nuup kangerluani takuneqartarsimasut takuffiup eqqaalu aqqusaarsimagaluarlugu takuffiginagillu. Ulluni taakkunani kisitsinittinni suleqatigiilluarluta pisimavarput silagissuup ataani. Nuup kujataatungaa eqqaalaarusullugu ukiuunerani tuttunik takussaasuinnera maluginiarsimagakku Erseqqinnerusumillu oqaatigisinnaanagu tamanna aatsaat aappaagumut pineqassagunarmat. Tuttunik kisitsineq. Kangerlussuaq + Nassuttooq Ulloq 18/3-2005 miit. 22/03-2005 ilanngullugu.Tuttunik kisitsineq. Ukuulluta peqataalluta Rink Heinrich. Hans Mølgaard, taavani piniarnermut nakkutilliisoq. Maluginiakkatta ilagaat kangerlussuup eqqaa tuttunissimaqisoq kisitsinermi siullermut sanilliullugu Maluginiakkattalu ilagaat nunarujussuaq qanoq tummaarineqartigisimasoq tamarluinnangajammi arqusinnerluni . Maluginiakkamalu aamma ilagaat Tuttut Umimmaallu imminnut akornuteqanngitsumik nunaqqatigiissinnaasut paasillugu, naak siornatigut tusartakkagut tuttut umimmaallu imminnut sapertut kisiannili taamaannani. Ulluni taakkunani assuttaaq suleqatigilluarluta kisitsineq ingerlassimavarput , assullu uanga nammineq paasisaqarluarlunga. Uangalu immikkut Christinimut qujarusuppunga suleqatigiilluarnitsinnut ulluni taakkunani. Immikkut eqqaasaqalaarusuppunga kangerlussuup nunataa pillugu. Tuttut Umimmaallu neriniarfii asororluinnassanngippata immikkut inatsisiliornikkut iliuuseqartoqartariaqarpoq. Inuss. inuull. Rink Heinrich Appendix 9 Transect snow cover and visibility conditions Photographs taken during the helicopter survey, March 2005 A) Central region, 14-16 March 2005 High-density stratum, from seacoast to Ice Cap Figure 12. The southern portion of Akia (Nordlandet), looking from Godhåbsfjord across Akia to the Davis Strait. Snow conditions typical for transects 77, 97, 17 and 39. Photo by: C. Cuyler. Figure 13. Akia, across from Nuuk, a closeup of the Kanasut uplands. Photo by: C. Cuyler. 1 Central region: High-density stratum, from seacoast to Ice Cap Figure 14. Akia, looking west-northwest from Kanasut. Photo by: C. Cuyler. Figure 15. Akia, looking west-northwest aross the Sârdlup taserssua Lake. Snow conditions typical for transect 68. Photo by: C. Cuyler. 2 Central region: High-density stratum, from seacoast to Ice Cap Figure 16. Akia, looking north into highlands. Photo by: C. Cuyler. Figure 17. Akia, west side of Quagssûp taserssua Lake, looking north to Usuk lake. Snow conditions typical for transect 227 and 18. Photo by: C. Cuyler. 3 Central region: High-density stratum, from seacoast to Ice Cap Figure 18. Akia, on the east side of the southern tip of Quagssûp taserssua Lake, 104 caribou were present in the foreground flatlands. Photo by: C. Cuyler. Figure 19. Transect 56 at the north-northwest end looking west at unnamed lake. Photo by: C. Cuyler. 4 Central region: High-density stratum, from seacoast to Ice Cap Figure 20. Looking east-northeast into Eldorado Valley No. 2, with Tugdluptâp tasia lake crossing the foreground. Snow conditions typical for transect 46. Photo by: C. Cuyler. Figure 21. Transect 65 area just east of Eldorado Valley No. 2. Photo by: C. Cuyler. 5 Central region: High-density stratum, from seacoast to Ice Cap Figure 22. Area to the north of Narssarssuaq Valley near transects 193 and 36. Photo by: C. Cuyler. Figure 23. Narssarssuaq Valley, southeast end of transect 8, looking north. Photo by: C. Cuyler. 6 Central region: High-density stratum, from seacoast to Ice Cap Figure 24. Narssarssuaq Valley, southeast end of transect 21, looking west-northwest. Photo by: C. Cuyler. Figure 25. Transect 108 at northeast end of Ilulialik Fjord. Photo by: C. Cuyler. 7 Central region: High-density stratum, from seacoast to Ice Cap Figure 26. Ilulialik Fjord, looking south out the fjord towards Bird Mountain. Photo by: C. Cuyler. Figure 27. Ilulialik Fjord, river valley on the east side. Photo by: C. Cuyler. 8 Central region: High-density stratum, from seacoast to Ice Cap Figure 28. Highlands near the Ice Cap in the area of transect 166. Photo by: C. Cuyler. Figure 29. Highlands near the Ice Cap at transect 64. Photo by: C. Cuyler. 9 North region: Low-density stratum, from seacoast to Ice Cap Figure 42. Transect 161 beginning at southwest end on lake. Photo by: C. Cuyler. Figure 43. Transect 151 beginning at southwest end, evidence of feeding craters. Photo by: C. Cuyler. 16 North region: Low-density stratum, from seacoast to Ice Cap Figure 44. Transect 155 looking northeast; snow cover typical of the seacoast areas near Nordre Strømfjord. Photo by: C. Cuyler. Figure 45. Transect 29 at the west-southwest end; snow cover typical of the mountain areas near Nordre Strømfjord. Photo by: C. Cuyler. 17 North region: Low-density stratum, from seacoast to Ice Cap Figure 46. Transect 27, start in northeast on left and southwest finish on right. Snow cover changed dramatically along the varied elevation gradient of many transects. Photo by: C. Cuyler. Figure 47. Transect 77, north of Søndre Strømfjord and midway between seacoast and Ice Cap, had little remaining snow cover. Photo by: C. Cuyler. 18 North region: Low-density stratum, from seacoast to Ice Cap Figure 48. Transect 64, south of Søndre Strømfjord and midway between seacoast and Ice Cap. Photo by: C. Cuyler. Figure 49. General lack of snow cover in the area southwest of transect 158. Photo by: C. Cuyler. 19 North region: Low-density stratum, from seacoast to Ice Cap Figure 50. Transect 61, area just south of the large Taserssuaq Lake, view of the east side of the Qangatap kua River Valley looking northwest towards the lake. Photo by: C. Cuyler. Figure 51. Transect 125 near the Ice Cap: taken while flying the transect in an east to west direction, looking south. Photo by: C. Cuyler. 20 North region: Low-density stratum, from seacoast to Ice Cap Figure 52. No snow close to the Ice Cap and near transect 125, looking north with Angmalortoq Lake in the foreground. Photo by: C. Cuyler. Figure 53. Area next to the Ice Cap (not shown but on immediate right), at the point of highest elevation on transect 32 looking northwest to Ørkandalen Valley. Photo by: C. Cuyler. 21 North region: High-density stratum, inland and to Ice Cap (before snowfall) Figure 54. North shore of Kangerlussuaq Fjord, looking west towards Amitsorssuaq Lake, and approaching general area near the western end of transect 24. Photo by: C. Cuyler. Figure 55. Transect 24 near the southwest end. Photo by: C. Cuyler. 22 North region: High-density stratum, inland and to Ice Cap (before snowfall) Figure 56. General area to the north of transect 183 looking east (far end of furthest lake centre is has DeHaviland airplane wreck). Photo by: C. Cuyler. Figure 57. Approaching the west-southwest end of transect 152, looking east with Sanningassoq Lake in the distance. Photo by: C. Cuyler. 23 North region: High-density stratum, inland and to Ice Cap (before snowfall) Figure 58. Flying transect 16 looking northwest. Photo by: C. Cuyler. Figure 59. Transect 193 in the northern portion of the high-density stratum, looking to the northwest from the southwest end of the transect. Photo by: C. Cuyler. 24 North region: High-density stratum, inland and to Ice Cap (after snowfall) Figure 60. Area between the Isortoq River and Kangerlussuaq airport, specifically west of transect 116 looking north-northwest after the light snowfall. The salt and pepper background made spotting caribou difficult. Photo by: C. Cuyler. Figure 61. Area between the Isortoq River and Kangerlussuaq airport, specifically just north of Langsøerne. Salt and pepper background made caribou counting difficult. Photo by: C. Cuyler. 25 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 74. Nagssugtûp Nunâ area, flying transect 209, illustrating filtered lighting and salt & pepper background. Photo by: C. Cuyler. Figure 75. Nagssugtûp Nunâ area, illustrating typical conditions of fog, poor lighting and salt & pepper background. Photo by: C. Cuyler. 32 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 76. Nagssugtûp Nunâ area, flying transect 189 from northwest to southest. Caribou were well camouflaged against the dark salt & pepper background. Photo by: C. Cuyler. Figure 77. Nagssugtûp Nunâ area, flying transect 189 northwest to southeast. Photo by: C. Cuyler. 33 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 78. Nagssugtûp Nunâ area, transect 197 at the northwest end, looking northwest. Caribou were well camouflaged against the salt & pepper background. Photo by: C. Cuyler. Figure 79. Nagssugtûp Nunâ areas near the Ice Cap while flying transect 59, looking northwest. In the poor light, the caribou were camouflaged in the salt & pepper background. Photo by: C. Cuyler. 34 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 80. Eqalungmiut Nunât area, looking northwest at north shore of the Kuuk River at Igssuit. Photo by: C. Cuyler. Figure 81. Eqalungmiut Nunât area, looking west at the northeast end of Eqalungmiut Lake. Photo by: C. Cuyler. 35 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 82. Eqalungmiut Nunât, area northeast of Eqalungmiut Lake. Caribou were well camouflaged against the salt & pepper background. Photo by: C. Cuyler. Figure 83. Eqalungmiut Nunât at the southeast end of transect 63. Caribou were well camouflaged against the salt & pepper background area. Photo by: C. Cuyler. 36 North region: High-density stratum, inland and to Ice Cap (poor conditions) Figure 84. Eqalungmiut Nunât, south shore of the Qordlortoq River, which was just south of transect 154. Caribou hid well in the salt & pepper background. Photo by: C. Cuyler. Figure 83. Eqalungmiut Nunât at transect 154, illustrating poor survey conditions in the northern portion of the high-density stratum. Photo by: C. Cuyler 37 Appendix 10 Caribou trails in the North region Extensive trail networks, eroding and trampling in the high-density stratum between the Isortoq River and the Kangerlussuaq airport. Photographs taken following a light snowfall, 21 March 2005. Figure 84. Erosion trails created by caribou. The trails generally run on an east-west orientation. Photo by: C. Cuyler. 38 Figure 85. Erosion trail networks created by caribou. The trails run east-west. Photo by: C. Cuyler. 39 Figure 86. Three photos in series illustrating a typical network of caribou erosion trails covering an extensive area. Photos by: C. Cuyler. 40 Appendix 11 Loess storms near the Ice Cap in the North region Photographs were taken during the helicopter survey 18-22 March 2005 Figure 87. Loess storm blowing down from the Ice Cap and through the Isortoq River valley. Photo by: C. Cuyler. Figure 88. Loess storm caused by strong winds blowing down from the Ice Cap and through the Sandflugt and Ørkandalen valleys. Photo by: C. Cuyler. 41