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.

Application of machine-learning methods in forest ecology: recent progress and future challenges

“…In this article, we also summarize several advantages and disadvantages of these methods (see Table 1). For more detailed information of other ML algorithms see previous reports (Haupt et al 2008;Hsieh 2009;Michalski et al 2013;Muhamedyev 2015;Thessen (2016).…”

“…Currently, not only ANN but also most ML algorithms are very complex. These algorithms typically require strong mathematical skills and major investments in time to understand them in detail (Thessen 2016) as well as to avoid "black box" and overfitting problems. Although the unfamiliarity with "black box" effects does not necessarily hamper the use of ML algorithms, it may influence which algorithms are selected by users.…”

“…Previous reviews and books (Haupt et al 2008;Hsieh 2009;Thessen 2016) focused on several fields of research that included oceanography, hydrology, and atmospheric sciences, but they rarely reported on how ML was used to study forest ecosystems although these methods had increasingly become popular over the last decade in forest ecosystem research as reflected by the increasing number of publications (see Fig. 1).…”

“…1). In this research field, ML approaches provided powerful and efficient ways to deal with data that was nonlinear, had high dimensionality, contained complex interactions and/or missing values (Bhattacharya 2013;Thessen 2016). For example, using modern remote sensing and mapping techniques, ML methods effectively improved the accuracy of species distribution models (SDMs) (Garzón et D r a f t Vaca et al 2011;Pouteau et al 2012;Faleiro et al 2013;Périé and Blois 2016), or in combination with traditional processes or empirical models, they were used to predict carbon (C) and energy fluxes (Papale and Valentini 2003;Papale et al 2015;Cropper 2008, 2010;Tramontana et al 2015Tramontana et al , 2016.…”

Designing Autonomy: Opportunities for New Wildness in the Anthropocene

“…Thus the prospect of intelligent systems capable of acting autonomously in real time to sustain the autonomy of nonhuman species and ecological processes without direct human intervention appears increasingly possible. Indeed, are increasingly using machine-learning methods to develop species distribution models that inform conservation decisions [19][20][21]. Conservation biologists and managers are also employing deeplearning systems and other technologies to eliminate, counter, or mitigate anthropogenic influences on species and ecological processes (Box 1).…”