Comparing models using air and water temperature to forecast an aquatic invasive species response to climate change

Abstract

Understanding invasive species spread and projecting how distributions will respond to climate change is a central task for ecologists. Typically, current and projected air temperatures are used to forecast future distributions of invasive species based on climate matching in an ecological niche modeling approach. While this approach was originally developed for terrestrial species, it has also been widely applied to aquatic species even though aquatic species do not experience air temperatures directly. In the case of lakes, species respond to lake thermal regimes, which reflect the interaction of climate and lake attributes such as depth, size, and clarity. The result is that adjacent waterbodies can differ notably in thermal regime. Given these obvious limitations of modeling aquatic species distributions using climate data, we take advantage of recent advances in simulating lake thermal regimes to model the distributions of invasive spiny water flea (Bythotrephes cederströmii) for current and projected future climates in the upper Midwest of the USA. We compared predictions and future projections from models based on modeled air temperatures with models based on modeled water temperature. All models predicted that the number of suitable lakes in the region will decrease with climate change. Models based on air and water temperature differed dramatically in the extent of this decrease. The air temperature model predicted 89% of study lakes to be suitable, with suitability declining dramatically in the late century with climate warming to just a single suitable lake. Lake suitability predictions from the water temperature model declined to a much lesser degree with warming (42% of lakes were predicted to be suitable, declining to 19% in the late century) and were more spatially independent. Our results expose the limitations of using air temperatures to model habitat suitability for aquatic species, and our study further highlights the importance of understanding lake‐specific responses to climate when assessing aquatic species responses to climate change. While we project a contraction in the potential range of Bythotrephes with warming in the study region, we anticipate that Bythotrephes will likely continue to expand into new lakes that will remain suitable in the following decades.

Bat algorithm optimised extreme learning machine (Bat‐ELM): A novel approach for daily river water temperature modelling

“…In response to the increasing demand for T w data which nowadays are included in major environmental studies, especially climatic changes, continuous knowledge of its variation has becoming a priority for a large number of countries worldwide and has been included in their monitoring and management programmes (Tavares et al, 2020). For example, it was demonstrated that T w can be used as an important predictor for aquatic invasive species (Walsh et al, 2020), and it should be taken into account that T w should be considered in all studies and knowledge related to the habitat of river ecosystems (Morid et al, 2020). Although T w is one of the simplest measures of water quality, available information is still limited in terms of temporal frequency and representativeness (Yang et al, 2021).…”

Resist‐accept‐direct (RAD) considerations for climate change adaptation in fisheries: The Wisconsin experience

“…Beyond changes in habitat, losses of winter ice are yielding uncertain outcomes for ice fishing (Sharma et al, 2020), historically a period of high effort in Wisconsin (Feiner et al, 2020b). Climate change also threatens to exacerbate the risks of other stressors like heat‐induced fish kills (Till et al, 2019), range expansions of invasive species (Walsh et al, 2020), and more frequent and extreme floods (Dauwalter & Mitro, 2019; Gudmundsson et al, 2021), which can substantially threaten valuable native fish communities.…”

An Open-Source, Durable, and Low-Cost Alternative to Commercially Available Soil Temperature Data Loggers

“…Temperatures vary across time and space and can substantially differ between the air, water, and soils. Although air temperatures have been extensively measured using global networks or satellites, there has been less attention towards soil and water temperatures [ 3 , 6 , 9 , 10 ]. This is unfortunate because many biological and biogeochemical processes rely on proximate temperatures in soil and water that can differ widely from those in the air [ 10 , 11 , 12 , 13 , 14 , 15 ].…”

“…Although air temperatures have been extensively measured using global networks or satellites, there has been less attention towards soil and water temperatures [ 3 , 6 , 9 , 10 ]. This is unfortunate because many biological and biogeochemical processes rely on proximate temperatures in soil and water that can differ widely from those in the air [ 10 , 11 , 12 , 13 , 14 , 15 ]. In soils, temperatures can be difficult to predict from those in the air because of their dependence on radiation load, surface energy budget partitioning, soil depth, and soil thermal properties that vary with soil texture and moisture content [ 12 , 13 , 14 , 16 , 17 , 18 ].…”