Global maps of lake surface water temperatures reveal pitfalls of air‐for‐water substitutions in ecological prediction

Abstract

In modeling species distributions and population dynamics, spatially-interpolated climatic data are often used as proxies for real, on-the-ground measurements. For shallow freshwater systems, this practice may be problematic as interpolations used for surface waters are generated from terrestrial sensor networks measuring air temperatures. Using these may therefore bias statistical estimates of species' environmental tolerances or population projections -particularly among pleustonic and epilimnetic organisms. Using a global database of millions of daily satellite-derived lake surface water temperatures (LSWT), I trained machine learning models to correct for the correspondence between air and LSWT as a function of atmospheric and topographic predictors, resulting in the creation of monthly high-resolution global maps of air-LSWT offsets, corresponding uncertainty measures and derived LSWT-based bioclimatic layers for use by the scientific community. I then compared the performance of these LSWT layers and air temperature-based layers in population dynamic and ecological niche models (ENM). While generally high, the correspondence between air temperature and LSWT was quite variable and often nonlinear depending on the spatial context. These LSWT predictions were better able to capture the modeled population dynamics and geographic distributions of two common aquatic plant species. Further, ENM models trained with LSWT predictors more accurately captured lab-measured thermal response curves. I conclude that these predicted LSWT temperatures perform better than raw air temperatures when used for population projections and environmental niche modeling, and should be used by practitioners to derive more biologically-meaningful results. These global LSWT predictions and corresponding error estimates and bioclimatic layers have been made freely available to all researchers in a permanent archive.

Range expansion risk for a newly established invasive duckweed species in Europe and Canada

“…We retrieved occurrence records of dotted duckweed from the Global Biodiversity Information Facility (GBIF, https://www.gbif.org/) and from published papers (Appendix 3). The environmental predictors selected for modeling distribution were based on satellite-derived lake water temperatures, which better represent duckweed habitat conditions than the customarily used air temperature-derived variables (Armitage, 2023). Additionally, we compared Maxent models using air versus lake water temperatures.…”

“…Commonly used current air temperatures were obtained from the WorldClim database (https://www.worldclim.org; Merow, Smith, & Silander Jr, 2011; Fick & Hijmans, 2017). Lake water temperature predictors were calculated by Armitage (2023) and are based on GloboLakes, a global database of satellite-calculated daily temperatures of lake surfaces (www.laketemp.net/home_GL). The author adjusted the relationship between these lake temperatures and air temperatures using machine learning models based on various atmospheric and geographical factors.…”

“…We selected the bioclimatic predictors BIO 1 (mean annual temperature) and BIO 7 (annual range of temperature) to describe the thermal-dependency of dotted duckweeds. Such predictors were previously used in Maxent models describing the distribution of similar duckweed species, and are also known to be good predictors of duckweed thermal performance (see Armitage, 2023).…”

“…Optimal suitability estimated by the current air temperature model also occurs at much lower precipitation conditions in comparison with the other models (Table 5). Armitage (2023). This curve was then superimposed to BIO 1 response curves.…”

Optimization Driven Distributed Deep Learning for Aqua Status Prediction in IoT

“…In 2023, Armitage, D.W., [26] have adopted a generalized additive model (GAM) and ML. Initially, ML predicts the lake surface water temperatures (LSWT) and air in the atmosphere and generates a map globally with ambiguous measures.…”