The Upper Thames Water Vole Restoration Project

Success in water vole reintroductions

Reintroductions are a potentially important tool in conservation biology but frequently fail. Few studies have examined the factors that may influence the ‘success’ of a reintroduction.

Water voles (Arvicola terrestris) are an endangered species in the UK, with a population decline that has resulted from impacts of habitat loss and predation by American mink. The Upper Thames Water Vole Restoration Project ran from 2004 to 2008, aiming to re-establish populations of water voles  to areas of the Upper Thames from which they had previously been extirpated by American mink predation. This project was conceived as an experimental study, examining the effect of habitat quality upon reintroduction success, and examining the health and welfare implications of the reintroduction process for the individuals involved.

We selected 12 replicate 800 m stretches which supported different abundances of riparian vegetation (see Fig. 1). This vegetation is crucial for water voles as it forms not only their protection from predation, but also their main food source. The 12 sites were selected from within two categories: six sites had riparian vegetation that was 1-3 m in thickness, and six had riparian margins that were 3-6 m in thickness. In practise, sites represented a continuum of habitat thickness from 1- 6 metres. We established mink control (following the same methods as our mink removal experiment) at each reintroduction site sufficiently in advance of the release to remove American mink from the locality of the reintroduction. Mind control was maintained throughout each reintroduction and is now being continued by our project partners (BBOWT and the Lower Windrush Project).

Cohorts of 44 water voles with a 1:1 sex ratio were in released in each May over three years. Water voles initially established at nine sites (See Table 1), failing to establish at three sites due to predation from American mink (two sites) and atypically severe flooding post release (one site). Each case of failed mink control occurred at a site where the project relied upon landowner participation. For sites where voles established, at those with higher vegetation abundance more of the release cohort survived (initial survival rates range 0.43 - 0.61), and post-establishment survival rates (range 0.45 â€" 0.80) and population densities (range 2.1 - 5.4 voles per 100m of habitat) were higher. A further two populations were lost to American mink predation post establishment.

Reintroductions are commonly designated as either a ‘success’ or a ‘failure’. The principal cause of a failed release in our study was insufficient mink control. However, whilst seven of our twelve reintroductions were ‘successful’, our results indicated substantial variation in the population densities and survival rates that the replicate habitats could support. This highlights the need to ensure that any habitat selected for a reintroduction is the best obtainable.

Table 1 - Summary of the results of the reintroductions and causes of failure. Numbers in brackets represent the number of populations that were lost for a particular reason during the period. These releases took place over three years, but year is omitted for clarity.                                                                                                   

Stage of   reintroduction	End of month	Number of extant   populations	Reason for loss of   population(s)
Initial release	May	12	/
Establishment	June	9	Flooding post release   (1)       Failure of mink   control (2)
Survival to end of   breeding season	October	8	Failure of mink   control (1)
Successful   overwintering of population	April	7	Failure of mink   control (1)

Growth and maturation rates in water vole populations
The restoration project offered a unique opportunity to study the link between forage abundance, individual growth rates and maturation rates in small mammal populations. Studies examining the link between demographic rates and forage abundance in small mammal populations typically either use enclosure experiments or draw observations from extant populations. This project allowed us to examine this subject area with wild populations whilst controlling one factor: the amount of forage available to the populations.

In populations of small mammals, food supplementation typically results in higher population densities, body weights, growth rates and reproductive rates. However, few studies have demonstrated a relationship between forage levels and demographic rates in wild populations in the absence of supplementation. We examined the association of levels of available forage with individual growth rates and time to sexual maturity in the eight successfully established reintroduced populations, and also in three naturally-occurring populations of water voles.

We found that range-sizes were smaller at sites with higher population densities, as expected from our previous work in this area. However, the amount of forage available to an individual covaried with range-size at each site, meaning that individuals with larger ranges had more forage available to them. Similarly individual growth rates of individuals with larger ranges were also larger than individuals with small ranges.

Water voles, in common with several species of small mammal, become sexually mature at a specific weight, not at a specific age. The weight at which water voles became sexually mature was 112 g for females and 115 g for males and did not vary between study sites. Differences in growth rates therefore translated into differences in the time taken to reach maturity between sites. In the reintroduced populations, mean days to maturity varied inversely with mean range-length. Females took seven days (18%, range 40 - 47 days) longer and males 5 days (13%, range 40 â€" 45 days) longer to reach breeding condition at the sites with the shortest mean range-lengths.
This study, we believe, presents the first evidence from wild populations that density dependence (a feedback loop where populations at high density may have reduced breeding rates) may operate at least in part through suppression of forage availability and therefore individual growth rates.