Abstract
Ecosystem Services (ES) are an established conceptual framework for attributing value to the benefits that nature provides to humans. As the promise of robust ES-driven management is put to the test, shortcomings in our ability to accurately measure, map, and value ES have surfaced. On the research side, mainstream methods for ES assessment still fall short of addressing the complex, multi-scale biophysical and socioeconomic dynamics inherent in ES provision, flow, and use. On the practitioner side, application of methods remains onerous due to data and model parameterization requirements. Further, it is increasingly clear that the dominant “one model fits all” paradigm is often ill-suited to address the diversity of real-world management situations that exist across the broad spectrum of coupled human-natural systems. This article introduces an integrated ES modeling methodology, named ARIES (ARtificial Intelligence for Ecosystem Services), which aims to introduce improvements on these fronts. To improve conceptual detail and representation of ES dynamics, it adopts a uniform conceptualization of ES that gives equal emphasis to their production, flow and use by society, while keeping model complexity low enough to enable rapid and inexpensive assessment in many contexts and for multiple services. To improve fit to diverse application contexts, the methodology is assisted by model integration technologies that allow assembly of customized models from a growing model base. By using computer learning and reasoning, model structure may be specialized for each application context without requiring costly expertise. In this article we discuss the founding principles of ARIES - both its innovative aspects for ES science and as an example of a new strategy to support more accurate decision making in diverse application contexts.
Highlights
The advantages of an ecosystem services (ES) view of coupled human-natural systems have been widely recognized in science, management and governance [1]
Once discussions on ES became mainstream - thanks largely to the seminal Millennium Ecosystem Assessment (MEA) [2] - and lessons from many individual case studies were learned [3], a first generation of integrated, multi-ES assessment methodologies and tools has been striving to meet the needs of an audience that cuts across the academic, governmental, NGO, and corporate sectors [4,5]
To provide examples of early ARIES outputs, we summarize key results from two contrasting case studies: water supply and quality in Eastern Madagascar and open space values in Washington State, USA
Summary
The advantages of an ecosystem services (ES) view of coupled human-natural systems have been widely recognized in science, management and governance [1]. Focusing on both the biophysical mechanisms of ES provision and the socioeconomic implications of their use can allow decision makers to directly link natural capital to the societies and economies that depend on it. Rapid assessment methods have come to command wide interest from all these communities [5]. Most early assessment studies [8,9] and some recent methods [1,10] infer ES values through production functions whose driving input is land cover type, alone or complemented by limited other structural information (e.g., vegetation type). Other methods [3] have proposed models of a more functional nature to more accurately represent the mechanistic underpinnings of ES dynamics [11,12,13,14]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
