Phenological responses of 215 moth species to interannual climate variation in the Pacific Northwest from 1895 through 2013

Abstract

Climate change has caused shifts in the phenology and distributions of many species but comparing responses across species is challenged by inconsistencies in the methodology and taxonomic and temporal scope of individual studies. Natural history collections offer a rich source of data for examining phenological shifts for a large number of species. We paired specimen records from Pacific Northwest insect collections to climate data to analyze the responses of 215 moth species to interannual climate variation over a period of 119 years (1895-2013) during which average annual temperatures have increased in the region. We quantified the effects of late winter/early spring temperatures, averaged annually across the region, on dates of occurrence of adults, taking into account the effects of elevation, latitude, and longitude. We assessed whether species-specific phenological responses varied with adult flight season and larval diet breadth. Collection dates were significantly earlier in warmer years for 36.3% of moth species, and later for 3.7%. Species exhibited an average phenological advance of 1.9 days/°C, but species-specific shifts ranged from an advance of 10.3 days/°C to a delay of 10.6 days/°C. More spring-flying species shifted their phenology than summer- or fall-flying species. These responses did not vary among groups defined by larval diet breadth. The highly variable phenological responses to climate change in Pacific Northwest moths agree with other studies on Lepidoptera and suggest that it will remain difficult to accurately forecast which species and ecological interactions are most likely to be affected by climate change. Our results also underscore the value of natural history collections as windows into long-term ecological trends.

Fig 1. Response curve of date of capture for an early-season specialist, Leptarctia californiae (Walker, 1855), to regional February-April temperature anomalies.

The response curve was calculated with the GLMM model slope and intercept of the relationship between temperature anomaly and log_e(date of capture), but it is plotted using a linear Y-axis for ease of interpretation. Negative February-April anomalies represent colder than average years while positive values represent warmer than average years. (a) Response curve plotted with elevation. Elevation was included as a random effect in our GLMM because higher elevation species tend to fly later in the season, explaining some variation in model slope. (b) Collection years plotted with response curve to show changes in phenological response through past 119 years.

Fig 2. Frequency distribution of phenological sensitivities (days/°C) of moth species to regional February-April temperature anomalies.

Zero indicates no shift in flight date, negative values indicate earlier capture dates and positive values indicate later capture dates in years with a 1°C deviation from zero. Mean sensitivity is 1.9 days/°C across all 215 species.

Fig 3. Comparisons of model slopes among phenological and dietary functional groups.

(a,b) Comparing average model slope for the relationship between temperature anomaly and log_e(date of capture) (inverted for ease of interpretation). (c,d) Comparing percent of significant slopes for the percentage of species flying significantly earlier in warmer years. Significant pairwise differences are indicated by differences in letters above the bars. See Methods for explanation of seasonality and diet breadth categories.