Adoption of machine learning techniques in Ecology and Earth Science

Abstract

The natural sciences, such as ecology and earth science, study complex interactions between biotic and abiotic systems in order to infer understanding and make predictions. Machine-learning-based methods have an advantage over traditional statistical methods in studying these systems because the former do not impose unrealistic assumptions (such as linearity), are capable of inferring missing data, and can reduce long-term expert annotation burden. Thus, a wider adoption of machine learning methods in ecology and earth science has the potential to greatly accelerate the pace and quality of science. Despite these advantages, machine learning techniques have not had wide spread adoption in ecology and earth science. This is largely due to 1) a lack of communication and collaboration between the machine learning research community and natural scientists, 2) a lack of easily accessible tools and services, and 3) the requirement for a robust training and test data set. These impediments can be overcome through financial support for collaborative work and the development of tools and services facilitating ML use. Natural scientists who have not yet used machine learning methods can be introduced to these techniques through Random Forest, a method that is easy to implement and performs well. This manuscript will 1) briefly describe several popular ML methods and their application to ecology and earth science, 2) discuss why ML methods are underutilized in natural science, and 3) propose solutions for barriers preventing wider ML adoption.

The unrealized potential of herbaria for global change biology

“…New machine learning tools and Bayesian techniques allow for the analyses of complex, hierarchically structured, and incomplete data sets and are suited to analyzing large, sparsely sampled data, all common features of collections data. Many of these methods are already well developed for modelling species distributions, e.g., maximum entropy, generalized linear and additive models, boosted regression trees, and random forest (Elith and Leathwick ), though they are not yet integrated into ecology more generally (Thessen ).…”

Performances of machine learning algorithms for mapping fractional cover of an invasive plant species in a dryland ecosystem

“…They have used a variety of environmental, bioclimatic, and/or earth observation data, and applying classification or regression methods. More recently, machine learning algorithms (MLAs) have gained high popularity in ecology and earth science because of their ability to model highly dimensional and non‐linear data with complex interactions and deal with data gaps (Thessen, ). Good performances of MLAs have been obtained in several fields, including remote sensing classifications (Mountrakis, Im, & Ogole, ) and species distribution modeling (Cutler et al, ; Elith & Leathwick, ).…”

A hierarchical path-segmentation movement ecology framework

“…At this point, it is worth noting that accelerometer data alone have been used to parse out relatively short time scale (order of 10 s) behavioral elements along movement paths using machine learning methods (Nathan et al 2012;Thessen 2016;Wang et al 2015). These behavioral elements, which almost certainly include several FuME steps, in reality are partial elements of more extensive short-duration CAMs that typically last tens of seconds to several minutes.…”

“…The temporal range is only suggestive and best applied to medium and large vertebrates. Also the image placements are not precise and some methods, such as machine learning (Thessen 2016), can be applied to data at any scale, but here are associated with the scale at which they are likely to be most useful. In addition, deep learning (useful for identifying different types of long-term patterns) is actually a subset of machine learning (where other machine techniques, such as random forests and support vector machines have been applied to accelerometer data; Nathan et al 2012;Fehlmann et al 2017).…”