Estimating Ungulate Recruitment and Growth Rates Using Age Ratios

T rends in population grow th can be m onitored w ith data for key vital rates w ithou t knowledge o f abundance. A lthough adult female survival has the highest elasticity for ungulate population dynamics, the m ore variable recruitm ent rates are com monly m onitored to track local variation in grow th rates. Specifically, recruitm ent is often measured using late w inter youngradult age ratios, though these age ratios are difficult to reliably interpret given the contribution o f m ultiple vital rates to annual ratios. W e show th a t the supplem entation o f age ratio data w ith concurrent radio-telem etry m onitoring o f adult female survival allows bo th retrospective estim ation o f empirical population grow th rates and the decom position o f recruitm ent-specific vital rates. W e dem onstrate the estim ation o f recruitm ent and population grow th rates for 1 w oodland caribou population using these m ethods, including elasticity and life-stage sim ulation analysis o f the relative contribution o f adult female survival and recruitm ent rates to variation in population growth. W e show, for this w oodland caribou population, tha t adult female survival and recruitm ent rates w ere nearly equivalent drivers o f population growth. W e recom m end the concurrent m onitoring o f adult female survival to reliably interpret age ratios w hen m anaging caribou and other ungulates. © 2011 T h e W ildlife Society. KEY W O R DS age ratios, matrix models, Rangifer tarandus caribou, recruitment, survival, ungulates, woodland caribou. M onitoring population trends is fundam ental to species conservation. A bundance estimates are ideal for m onitoring population dynamics, hu t are often challenging to reliably obtain for species tha t are difficult to count due to rarity, uneven distributions, or poor detectability. Because knowl­ edge o f population trend may he as valuable for conservation as abundance per se, and estimates o f trend do not necessarily require estimates o f abundance, trend estimates are a com ­ m on surrogate for abundance in conservation planning. D em ographic models allow the estim ation o f population grow th rates using vital rates w ithou t requiring knowledge o f abundance, and in closed populations key vital rates can reduce to survival and recruitm ent (H a tte r and Bergerud 1991, D anchin et al. 1995, W h ite and Bartm ann 1998, Caswell 2001). Recruitm ent is com monly m onitored w ith youngradult age ratios derived from harvest and survey data, particularly for populations o f birds (Kaminski and G luesing 1987, M enu et al. 2002) and ungulates (Roseherry and W o o lf 1991, M onello et al. 2001, W h ite et al. 2001, B right and H ervert 2005, H egel et al. 2010). Several authors have cautioned th a t age ratios alone are not sufficient for populaReceived: 22 July 2010; Accepted: 30 April 2011; Published: 23 September 2011 ^E-mail: nkk.decesare@umontana. edu tion m onitoring (Caughley 1974, M cC ullough 1994, W hite and B artm ann 1998), though they have been shown to he positively correlated w ith population grow th rates for elk under some conditions (H arris et al. 2008). W e consider age ratios collected near the end o f the biological year (e.g., late w inter for ungulates th a t b irth in spring), w hen juveniles are considered as recruited to the adult population. These ratios are difficult to in terpret w ith specific reference to population grow th or vital rates given th a t they can he influenced by m ultiple contributing vital rates. T h e num erator, or num ber o f young at the tim e o f survey, is the product o f the agespecific fecundity o f adult females (which is itself the product o f pregnancy rate, fetal survival, and litter size) and the survival o f young to the tim e o f survey, whereas the denom i­ nator, or num ber o f adults at the tim e o f survey, is affected by the age-specific survival rates o f adults since the b irth pulse. T h e interpretation o f ungulate recruitm ent rates using age ratios is further com plicated by an unknow n age structure o f breedingand nonhreeding-aged adults, given a m inim um age o f first reproduction o f 2-years-old. T rained observers may not he able to reliably differentiate nonhreeding year­ lings (1 to 2-year-olds) and breeding-aged adults (>2-yearolds) in the total count o f adult females, especially during aerial surveys (Sm ith and M cD onald 2002). D espite these complications in interpreting age ratios, H arris et al. (2008) found tha t 96% o f the variation in 144 T h e Jou rna l o f W ild life M an ag e m en t • 76(1) sim ulated age ratios for elk {Cervus elaphus) was explained by variation in a single vital rate, calf survival. Consequently, in populations w here calf survival is the vital rate m ost predictive o f population grow th rates (GaiUard et al. 2000, Raithel et al. 2007), age ratios alone may be suitable for m onitoring population trends. For long-lived, iteroparous species such as ungulates, the elasticity o f population grow th rates is typically highest in adult female survival (Gaillard et al. 2000), bu t variation in this im portant vital rate is expected to be m inim ized by evolutionary adaptation (Pfister 1998, Gaillard and Yoccoz 2003). T he survival, or recruitm ent, o f young generally has higher variance than th a t o f adults (Gaillard et al. 1998), w hich affords this vital rate subsequently greater correlation to realized population grow th rates (Gaillard et al. 2000, Raithel et al. 2007). However, additional dem ographic analyses have revealed th a t th e variance o f vital rates and their resultant im portance to population grow th rates can vary across populations w ith in species and across tim e or space w ith in populations (Albon et al. 2000, M orrison and H ik 2 0 0 7 , Ezard et al. 2008, Nilsen et al. 2009, Johnson et al. 2010). Furtherm ore, the im portance o f adult survival in explaining ungulate population grow th rates may be accentuated in declining or endangered populations experiencing patterns o f adult survival th a t are lower or m ore variable than would have been evolutionarily stable (O w en-S m ith and M ason 2005, N ilsen et al. 2009, Johnson et al. 2010). T his highlights the potential im portance o f m onitoring bo th recruitm ent and adult survival for complete depiction o f population trend. W e show th a t concurrent m onitoring o f adult survival using radio-telem etry supple­ m ents age ratio data sufficiently to alleviate the initial con­ cerns o f Caughley (1974) and M cC ullough (1994) and allows the estim ation o f bo th recruitm ent and population grow th rates. A nnual population estimates o f a threatened ungulate, w oodland caribou {Rangifer tarandus cariboii), would inform recovery planning under Canada’s Species at Risk A ct (SARA), bu t m onitoring o f w oodland caribou abundance is difficult due to their low density, small group sizes, and low sightability in densely forested habitat. In A lberta and other jurisdictions, m onitoring o f woodland caribou has focused on the m onitoring o f adult female survival and age ratios, from w hich population grow th rates have been estim ated (M cLoughlin et al. 2003). W e use m onitoring data from a population o f w oodland caribou in A lberta as a case study for developing dem ographic models o f population grow th w ith adult female survival and age ratio data. W e dem onstrate the estim ation o f population grow th w ith 2 m ethods. H atte r and Bergerud’s (1991) B J M equation and m atrix population models. F or use w ith m atrix models, we show the proper adjustm ent o f age ratio data to isolate adult female recruitm ent rates. Lastly, we use adult survival, age ratio, and age o f first reproduction data to conduct a life-stage sim ulation analysis (LSA; W isdom et al. 2000), assessing the relative im portance o f each vital rate to population grow th for a single w oodland caribou population. STUDY AREA W e studied vital rates and population grow th for the A la Peche w oodland caribou population in the Rocky M ountains o f w est-central A lberta. T h e A la Peche population trad i­ tionally m igrates between an alpine sum m er range in pro­ tected areas 0asper N ational P ark and the W illm ore W ilderness Area) and a forested foothill w inter range east o f park boundaries, though individuals can exhibit sedentary behavior in bo th w inter and sum m er ranges (M cD evitt et al. 2009). T h e eastern portion o f the range has been m odified by industrial land use such th a t 59% o f the population’s range outside o f protected areas is w ith in 500 m o f an an thropo­ genic feature (e.g., road, seismic line, or forest harvest unit; A lberta Sustainable Resource D evelopm ent and A lberta Conservation Association [A SR D and A C A ] 2010). W o lf {Canis lupus) population control was conducted to benefit a neighboring caribou population in an area overlapping the eastern boundary o f the A la Peche w inter range during 2006-2009 (A SR D and A C A 2010). Caribou are protected from hunting w ith in the national parks, and licensed hunting o f the A la Peche population, w hen outside o f the park, was discontinued in 1981. Forested habitats included upland lodgepole pine {Pinus contorta), spruce {Picea spp.) and aspen {Populus tremuloides) mixed forests, and lowland black spruce {Picea mariand) muskegs.