Data bias in the Atlantic Rainforest: uncovering the elephant in the room

Suivi de la biodiversité
jeudi 19 déc. 09:45 AM (15 minutes)
Pause café   10:00 AM à 10:10 AM (10 minutes)
Salle AB

Following the technological development that enabled ecologists to explore ideas in large temporal and spatial scales, biodiversity data (such as taxonomic, geographical, functional and genetic) available online fill a need for representative data in similar scales, even when ecological, social and economic restrictions prevented data collection. However, the quantity of these data does not reflect its quality. As data collection is naturally susceptible to gaps, highlighting, mapping and analyzing its flaws allow us to build more accurate hypotheses, ask the right questions and advance our knowledge in innovative directions. Tetrapods inventory data is frequently used in ecological research as inputs, but its biases are rarely considered. Some of them can be related to environmental variables that are proxies for biodiversity, such as temperature, potential evapotranspiration (PET), topography and vegetation. I assessed if these variables are correlated to Tetrapods inventory incompleteness at the Atlantic Rainforest (AF) by mapping rarefaction curve slopes. Species richness is more related to temperature (R = 0.19, p < 0.01) and to topography (R = -0.18, p < 0.01), but among well sampled cells topography has an opposite effect (R = 0.38, p < 0.01) and PET becomes very important (R = -0.19, p < 0.01). Only 31.47% of this ecoregion could be considered well sampled, mostly near large preserved areas. This is a clear example of how biased data can lead to erroneous hypotheses, and how data gaps must cease to be the elephant in the room and start to be part of the ecological analysis.

Mon horaire

Ajouter à Mon Horaire comme Favoris