Simple signals indicate which period of the annual cycle drives declines in seasonal populations

Abstract

For declining wild populations, a critical aspect of effective conservation is understanding when and where the causes of decline occur. The primary drivers of decline in migratory and seasonal populations can often be attributed to a specific period of the year. However, generic, broadly applicable indicators of these season‐specific drivers of population decline remain elusive. We used a multi‐generation experiment to investigate whether habitat loss in either the breeding or non‐breeding period generated distinct signatures of population decline. When breeding habitat was reduced, population size remained relatively stable for several generations, before declining precipitously. When non‐breeding habitat was reduced, between‐season variation in population counts increased relative to control populations, and non‐breeding population size declined steadily. Changes in seasonal vital rates and other indicators were predicted by the season in which habitat loss treatment occurred. Per capita reproductive output increased when non‐breeding habitat was reduced and decreased with breeding habitat reduction, whereas per capita non‐breeding survival showed the opposite trends. Our results reveal how simple signals inherent in counts and demographics of declining populations can indicate which period of the annual cycle is driving declines.

Early warning indicators of population collapse in a seasonal environment

“…Data were collected as part of the same experiment described in Burant et al. (2019), in which replicate populations of Drosophila melanogaster were established with distinct breeding and non‐breeding periods. Seasonal populations were generated from an outbred, aseasonal stock population maintained at 25°C, 40% RH and a 12:12 hr light–dark cycle, with a 2‐week generation time.…”

“…During the breeding period, populations of flies were fed a yeast‐agar medium (see recipe in Burant et al., 2019). This medium, which is the standard substrate for maintaining flies in cage culture, provided not only a source of nutrients for breeding adults but also acted as a substrate on which females deposited their egg and in which larvae developed.…”

“…a population may experience chronic environmental degradation during the breeding period that is not mirrored by similar change during the non‐breeding period, or vice versa ; Calvert et al., 2009). Such seasonality in environmental stressors can make it difficult to identify when and where populations are being driven to decline, although recent evidence suggests that simple demographic signals may reflect the season in which the stressor occurs (Burant et al., 2019). Whether the response of early warning indicators differs depending on the period of the annual cycle in which the stressor occurs remains to be investigated.…”

“…(2019) showed that the pattern of population decline and change in demographic vital rates were predicted by the season in which habitat loss occurred, suggesting that vital rates themselves may be useful in identifying the season driving a population to decline. Here, we hypothesized that whether a suite of indicators acted as early warning signals of an impending population collapse would be dependent on the season of habitat loss because the capacity of populations to respond to environmental change and the way in which populations decline differs depending on the season in which habitat is lost (Burant et al., 2019). When non‐breeding habitat is lost, we hypothesized that populations would rebound the following breeding period as a result of density‐dependent reproduction, and that this rebound would dampen the production of signals in breeding time series.…”

“…When non‐breeding habitat is lost, we hypothesized that populations would rebound the following breeding period as a result of density‐dependent reproduction, and that this rebound would dampen the production of signals in breeding time series. In contrast, when breeding habitat is lost, there is no similar compensatory mechanism by which populations can rebound since no reproduction occurs during the non‐breeding period (Burant et al., 2019). Following this rationale, we predicted that, when habitat was lost during the non‐breeding period, the abundance and trait time series would produce fewer and less sensitive early warning signals compared to when habitat was lost during the breeding period.…”