Ecosystem services are essential for human well-being, but are currently facing many natural and anthropogenic threats. Modeling and mapping ecosystem services helps us mitigate, adapt to, and manage these pressures, but overall the field faces multiple major limitations. These include: 1) data availability, 2) understanding, estimation, and reporting of uncertainties, and 3) connecting socio-ecological aspects of ecosystem services. Recent technological advancements in machine learning coupled with rising availability of big data, offer an opportunity to overcome these challenges. We review studies utilizing machine learning and/or big data to overcome these limitations. We collect 56 papers that exemplify the current use of machine learning and big data to address the three identified gaps in the ecosystem service field. We find that although the use of these tools in ecosystem service research is relatively new, it is growing quickly. Big data can directly address data gaps, especially as new big data resources relevant to ecosystem service mapping become available (ex. social media data). Some properties of machine learning can also contribute to addressing data gaps in data sparse environments. Also, many machine learning algorithms can estimate and consider uncertainty, whereas big data can significantly increase sample size, reducing uncertainties in some situations. Some big data sources, like crowdsourced data, provide direct sources of social behaviors and preferences that relate to ecosystem service demand, thus allowing researchers to connect social and biophysical aspects of ecosystem services. Machine learning algorithms provide an effective and efficient tool for handling these large nonlinear socio-ecological datasets in tandem, giving researchers the ability to more realistically model and map ecosystem services without relying on oversimplified proxies or linear algorithms. Despite these opportunities, implementation is still lacking and limitations still hinder use.

Understanding the economic value of ecosystem services is necessary to facilitate sustainable land use management, and to inform policy and decision making. However, valuing and monetizing ecosystem services remains challenging. Benefit-transfer and non-market valuation methods typically rely on administrative data and surveys, but this is time consuming, limited, and requires much more resources. Social media and other types of big data provide accessible and georeferenced data that can be incorporated into valuation approaches. We use recreation as an example and the Tahoe Central Sierra Initiative (TCSI) project area in California as a case study to explore the usefulness of such data in estimating travel costs that form an integral part of determining the value of recreational ecosystem services through the travel cost model. We estimated 6,951 person user days of recreation from 2,245 visitors who uploaded photographs to the Flickr photo-sharing application between 2005 and 2019. We used metadata from the images to infer visitor origins and estimate trip distance and costs of travel for visitors that took day trips (<500 miles (∼800 kms) roundtrip) to the area. Our results show that the most demand for recreational opportunities in the TCSI came from domestic visitors, particularly those from California and Nevada who took day trips. On average, visitors spent $156 per single day trip. The total cost of travel for recreational visits to the TCSI for the period was $491,500 (an average of $32,800 per year). However, when adjusted to align with actual visitation, the travel costs could range from $1.35 to $1.84 billion per year. Estimating recreational use and highlighting the travel cost for recreational opportunities illustrates how crowdsourced data can refine valuation approaches such as the widely used travel cost approach, which may fill in data gaps in valuing ecosystem services.

Crowdsourced geotagged social media data and machine learning approaches have emerged as promising tools for mapping ecosystem services, especially cultural ecosystem services that are difficult to assess. Here, we use recreation to show how social media data, machine learning, and spatial analysis techniques can improve our understanding of human-nature interactions and the mapping of recreational ecosystem services. We extracted 80,500 photographs taken in non-urban areas of the Tahoe Central Sierra Initiative project area in California between 2005 and 2019 that were posted to the photo sharing application Flickr and used these as a proxy for recreational visits to the area. Automated image content analysis was used to identify the objects and concepts in the photographs and uncover the types of nature experiences that are important to visitors. Additionally, variable importance, a Random Forest machine learning technique, was used to examine the environmental and landscape variables that drive recreation in the area and to create a classification model that predicts the recreation potential of the entire area based on important variables. The automated image content analysis identified 1,239 unique labels linked to recreation, with mountains, hills, and rocks being the most prominent features (22%). Our Random Forest model indicates that vegetation cover, land cover, elevation, smoke days, and landscape features are major drivers of recreation in the area and are of interest to visitors in the area. The model predicted that 25.9% of the area has the potential to support recreational visits. Most of these recreation potential areas are in protected areas (77.8%), predominantly in conifer forests (66%) and within national forest boundaries, especially the Tahoe National Forest area (37.6%). These results show that recreational ecosystem services vary across landscapes and illustrate the need for improved mapping approaches to determine the provision of ecosystem services in different places. The analysis provides novel insights into the various ways social media data and machine learning techniques can be powerful components of ecosystem service research and how they hold great potential for monitoring and informing management interventions on ecosystem service provision, especially in places with limited traditional onsite visitation data.

Forest restoration through mechanical thinning, prescribed burning, and other management actions is vital to improving forest resilience to fire and drought across the Western United States, and yields benefits that can be monetized, including improvements in water supply and hydropower. Using California's Sierra Nevada as a study area, we assess the water and energy benefits of forest-restoration projects. By using a scalable top-down approach to track annual evapotranspiration following forest disturbance, coupled with hydropower simulations that include energy-price information, and marginal prices for water sales, we project the potential economic benefits of hydropower and water sales accruing to water-rights holders. The results found that water-related benefits from strategically planned fuels-reduction treatments now being carried out can be sufficient to offset costs of management actions aimed at forest restoration, especially in the face of climate change. Our findings justified investments in restoring forests and reinforce the central role of water and hydropower providers in partnerships for management of source-water watersheds. Results also highlighted the importance of accurate, scalable data and tools from the hydrology and water-resources community.

Forests across the Western U.S. face unprecedented risk due to historic fire exclusion, environmental degradation, and climate change. Forest management activities like ecological thinning, prescribed burning, and meadow restoration can improve landscape resilience. Resilient forests are at a lower risk of high-intensity wildfires, drought, insects, and other disturbances and provide a wide range of benefits to ecosystems and communities. However, insufficient funding limits implementation of critically needed management. To address this challenge, we propose a multi-benefit framework that leverages the diverse benefits of forest management to engage a suite of stakeholders in sharing project costs. We take a three-pronged approach to develop our conceptual model: examining existing frameworks for environmental project implementation, conducting a literature review of forest management benefits, and evaluating case studies. Through our framework, we describe the steps to engage partners, starting by identifying benefits that could accrue to potential public and private beneficiaries, and moving through an iterative and collaborative process of valuing benefits, which can accrue over different spatial and temporal scales, in close consultation with potential beneficiaries themselves. The aim of this approach is to stack funding streams associated with each valued benefit to fully fund a given forest management project. The multi-benefit framework has the potential to unlock new sources of funding to meet the exceptional challenges of climate and wildfire disturbances. We apply the framework to dry forests of the Western U.S., but opportunities exist for expanding and modifying this approach to any geography or ecosystem where management provides multiple benefits.