| Abstract Detail
Biodiversity Informatics & Herbarium Digitization Hearn, David [1]. Sampling biases in herbarium data inflate signatures of climate change: causes and solutions. Hundreds of millions of plant specimens records are now available due to the efforts to digitize herbaria. These data provide a global, multi-century view of plant phenological change in response to climate and landscape alterations. Climate warming scenarios predict earlier onset of phenophases in more recent years. Despite the wealth of data, widespread sampling biases may obscure relationships among phenology, climate, and geography. Analyses of over 10,000 occurrence records from 13 NE North American spring ephemerals revealed confounding biases in latitudinal sampling and systematic shifts in sampling effort over time. Simulations suggested that these two sources of bias are among the most confounding. Methods to correct for these biases based on the theory of order statistics were developed. Before correcting for biases, 10 species showed a significant shift toward earlier phenophases in the year, whereas six (fewer than half) maintained a significant shift after corrections for biases. Before correcting for biases, all but one of the species showed a significant shift toward an earlier first date of phenophase onset. Only one of these shifts remained significant after corrections. Collectively, these results indicate that two widespread sources of bias artificially inflate signatures of climate change. Hence, there is a critical need to assess and contend with sampling biases when using herbarium specimens to analyze historical patterns of phenological change. The quantitative methods developed herein correct for these issues.
1 - Towson University, Biology, 8000 York Road, Towson, MD, 21252, United States
Keywords:
sampling bias
climate change
Herbarium Digitization
detrending
order statistics
phenology
spring ephemeral.
Presentation Type: Oral Paper
Number: BI&HD II006
Abstract ID:384
Candidate for Awards:None |
