Introduction
For species living in seasonal environments (e.g. from summer to winter or from dry to wet season) local adaptations to annually changing environmental conditions may evolve. Numerous species have evolved to time their life history events to match these changes in local seasonal conditions (Bradshaw and Holzapfel 2007, Williams et al. 2015). To time these phenological events, one reliable “zeitgeber” is daylength or photoperiod (Gwinner 2003, Hofman 2004), either on its own or in combination with other variables, such as temperature (Watson 1963, Jackes and Watson 1975, Larkin et al. 2001) and snow cover (Watson 1963, Flux 1970, Zimova et al. 2014).
Many animal species use photoperiod to time breeding (Goldman 1991, Gwinner 1996b, Dawson et al. 2001, Coppack and Pulido 2004), moulting (Lesher and Kendeigh 1941, Lyman 1943, Bissonnette and Bailey 1944), and migration (Gwinner 1996a), and other life history events. As photoperiod remains constant between years at specific locations, between year variation in local conditions could result in photoperiod timed phenological events being mistimed against the local environment. Fluctuations in environmental variables, such as precipitation (Villellas et al. 2014) and temperature (Ashmore and Janzen 2003, Kreyling et al. 2019), can result in increased within-species phenotypic variation in a variety of plant and animals species when compared to individuals of the same species that live in more stable habitats. Consequently, synchrony in phenological timing of individuals within a population is expected to increase with climate stability.
Animals occupying areas that are seasonally covered by snow live in environments that change from dark in summer to white in winter. As a predator avoidance strategy, at least 21 species (Mills et al. 2018, Zimova et al. 2018) have adapted seasonal changes in colouration of fur and feather, which provides camouflage in both a winter white and summer dark landscapes (Wallace 1879, Cott 1940, Merilaita and Lind 2005). To provide optimal camouflage, the timing of coat colour change should be synchronised with the period of continuous snow cover. Individuals would thus be expected to adapt the timing of their coat colour change to local conditions. Indeed, mismatched timing of coat colour change is linked to range contractions and population declines in several species including snowshoe hares, (Diefenbach et al. 2016, Sultaire et al. 2016), mountain hares (Acevedo et al. 2012, Pedersen et al. 2017), rock ptarmigan (Imperio et al. 2013), and white-tailed ptarmigan (Wang et al. 2002), showing the importance of correct timing. However, snow conditions might not be stable from year to year, and there might be seasonal differences in the predictability of the appearance and disappearance of snow.
There are some suggestions that snow cover might be more stable in spring compared to autumn. Snowshoe hares (Lepus americanus ) (Zimova et al. 2014) and least weasels (Mustela nivalis nivalis ) (Atmeh et al. 2018) exhibit limited phenotypic variation in moult timing in parts of their distribution during the spring moult, when transitioning from white to brown, but not during the autumn moult, when transitioning from brown to white. Therefore, the timing of seasonal coat colour change is expected to be more synchronised in spring compared to autumn. This could result in clearer differences in the timing of among years moult timing in the spring compared to the autumn. However, this is the first long-term study over a geographical area large enough to test these predictions.
Mountain hares (Lepus timidus ) express seasonal coat colour change in most of their range, except the subspecies of Irish hare (L. t. hibernicus ) found in Ireland (Mills et al. 2018). They are a generalist herbivore inhabiting boreal and alpine areas that occupy a wide range of climatic, latitudinal, and altitudinal gradients, experiencing large variations in winter snow cover duration. Coastal areas in the south and south-west of Norway experience relatively short snow cover duration compared to inland areas and areas in the north (Schuler et al. 2006) with coastal areas in the south and south-west receiving as little as one month of snow cover per year (Tallaksen et al. 2018). Additionally, coastal areas experience greater between-year variation in the depth and extent of snow cover than inland areas (www.senorge.no).
Here, we provide the first quantitative study of variation in mountain hare moult timing with nine years of data using 678 camera locations along an extensive geographic gradient in Norway. We studied 1) how the timing of moult varied with local geographical conditions and 2) how the timing of moult varied among years and seasons, utilising camera trap data in a Bayesian multinomial logistic regression model framework. Snowshoe hares that live at lower altitudes and latitudes displayed winter coats for a longer time compared to their low latitude (Grange 1932, but see Zimova et al. 2019) and altitude conspecifics (Holmgren et al. 2001, Nowak et al. 2020, Zimova et al. 2020b). Also, increased snow cover in continental areas is likely to result in hares living in these areas keeping their winter coats for longer than hares in coastal areas. Therefore, we used altitude, latitude, and climatic zone, distinguishing between coastal and continental climates, as indicators of local geographical conditions for our first aim. For our second aim, we predicted that moult timing would be more synchronised among individuals in spring compared to autumn, based on previous studies on snowshoe hares (Mills et al. 2013, Zimova et al. 2014). Furthermore, we predict larger among year variation in the timing of moult in spring compared to autumn.