Using Technology and Analytics to Enhance Stakeholder Engagement in Environmental Decision-Making

Enacted over 50 years ago, the National Environmental Policy Act (NEPA) calls for public participation in the environmental review of proposed actions by U.S. federal agencies. Much has changed in that period from the explosion in technology to the increasing expectation of stakeholders from participation to engagement. 

A recent IBM Center report, Using Technology and Analytics to Enhance Stakeholder Engagement in Environmental Decision-Making, by Jenna Yeager of the Public Lands Foundation, explores how stakeholder engagement in federal environmental decision-making has been facilitated through use of technology and how federal land management agency tools could be enhanced using analytics and artificial intelligence (AI).

This report covers NEPA activities as conducted by the four major federal land management agencies: Bureau of Land Management (BLM), National Park Service (NPS), and Fish and Wildlife Service (FWS) within the Department of the Interior; and the Forest Service (FS) within the Department of Agriculture, specifically in stakeholder engagement. Together, these four agencies manage over 650 million acres—a sizable portion of America’s public lands.

The key automated systems (‘apps’) from each agency that have external-facing components are:

• BLM: ePlanning/National NEPA Register
• NPS: Planning, Environment and Public Comment (PEPC)
• FWS: Environmental Conservation Online System (ECOS)/Information for Planning and Consultation (IPaC
• FS: Schedule of Proposed Actions (SOPA)

Yeager views the NEPA process through the lens of data and analytics rather than through the more common lens of governmental policy or organizational effectiveness. The research sought to identify what type of functionality external stakeholders need/expect when participating in agency NEPA actions.

Not surprisingly, according to the author, the existing stakeholder requirement frameworks were found to concentrate mostly on facilitation, trust building, communication, and similar soft skills—which are critical to the success of any endeavor—but were largely silent on tools/technology.

Developing a NEPA Stakeholder Engagement Framework

For that reason, Yeager creates a custom framework containing nine stakeholder requirements, rolled up into three categories:

• Discovering content: Stakeholders can review the proposed action, access agency content pertinent to the proposed action (‘what’ content is searched), search agency content pertinent to the proposed action (‘how’ text and spatial content is searched), and download data sets, analytical results, spatial data, APIs, etc. related to the proposed action.
• Analysis and context: Stakeholders can use interactive tools to explore the proposed action and assist in agency resource analysis performed under NEPA.
• Communication and collaboration: Stakeholders can comment on or submit alternatives for the proposed action, provide data and analysis directly to agency, and engage in communications with agency re: resource issues, decisions, and pre-or post-decision monitoring.

Technology for Enhancing NEPA Stakeholder Engagement

The report identifies key types of technology that have the potential of enhancing NEPA stakeholder engagement. For the most part, these new functionalities can be embedded into or interface with existing apps working in the background. only. Here’s preview of some of that technology under each of the three categories noted above.

Discovering content

• Cognitive (‘smart’) search is the new generation of information gathering technology. Using artificial intelligence (AI) capabilities such as natural language processing and machine learning to ingest, understand, and query digital content from multiple data sources, users receive results that are more relevant to their intentions.  If the agency provides ‘smart’ search capabilities, stakeholders can explore additional content (data, documents, images, etc.) that they consider useful for their evaluation, such as previous resource decisions informed by NEPA analysis, monitoring results, etc. Optimally, cognitive search functionality would be enabled through an external, user-friendly search interface (e.g., ‘show me evaluations over the past five years covering the outcome of sage grouse habitat restoration in Southeast Idaho’).

Analysis and context

• Business intelligence (BI) and data visualization tools in NEPA analysis would provide an additional, richer dimension for stakeholders to understand the proposed action not simply through words or static maps (largely the case now), but through graphics, charts, animations, dashboards, videos, and similar presentation vehicles. To date, these tools have seldom been used for stakeholder facing NEPA activities in the four agencies. The reasons vary from agency to agency, but usually involve staffing/funding shortages; just trying to address the significant NEPA workload often leaves little time to explore new possibilities.
• Text analysis is designed to derive value from text data when it is no longer humanly feasible to manually review and categorize that content. Stakeholders benefit from the agencies becoming timelier in their analyses/feedback. In addition, use of AI-assisted text analytics has the potential for better tracking and interpreting comments, even identifying issues that human interpreters might miss. In the interviews with BLM and the Forest Service, both agencies mentioned that they are contemplating or have undertaken small proof-of-concept projects that use text analytics.
• Sentiment analysis is contextual mining of text which identifies and extracts subjective information in source material. It gives agencies yet another communication channel. This type of analysis may be valuable at any stage, such as during pre-scoping (what issues are being raised?), during comment analysis periods (another channel the agency can tap into), or during alternatives analysis. In this manner, stakeholder feedback can be captured and interpreted, even though stakeholders may not be actively involved. Interview results did not show any of the four agencies using sentiment analysis as part of their NEPA analysis, but (anecdotally) it appears that sentiment analysis is conducted in other program areas.
• Location intelligence is the ability to derive business insights from geospatial information. Resource decisions informed by NEPA analysis are inherently spatial. They affect an area on the landscape. Location intelligence expands upon the functionality described above (BI visualizations) by providing the geospatial component in making resource allocation decisions—e.g., how the various natural features interact with or impact each other under different scenarios.All four agencies have a long history of using geospatial tools for solving business needs; that progress has continued as the discipline of location intelligence has transitioned into what many are now calling GeoAI.

Communication and collaboration

• Stakeholder engagement software is used by organizations to analyze their stakeholders, to create communication and engagement plans, to log information about the interactions they have with communities and to ensure compliance with regulations. Greater stakeholder trust/confidence in agency, seeing that their feedback has been received, acknowledged, and incorporated into the decision- making process. Also, greater trust from seeing that agency is following a defined, transparent process, and tracking/responding to communications made at any time, such as following designated comment periods.
• Data preparation tools, data integration tools, and master data management solution with AI/analytics working in the background can streamline the process of ren­dering imported data. Adoption of ‘newer’ agency data wrangling tools, many with AI in the back office, will significantly deepen stakeholder engagement, mov­ing from simply written comments to actual data—be it in the form of observations, data sets (including geospatial), or analytical results. Current agency NEPA policy encourages this level of engagement; the newer tools make it easier.

The research did not find any specific instances of AI and analytics currently being applied directly to NEPA stakeholder engagement. That’s not surprising, considering that AI (and, to some extent, advanced analytics) are relatively new—and being applied first in other management areas and disciplines. As their use in NEPA (and specifically stakeholder engagement) begins, best practices and policies will be developed.

Charting a Path Forward

Using the approaches outlined in this report, individual agencies can examine the maturity of stakeholder engagement and identify the key objectives they wish to address: what specific problems do they wish to solve in the near term or in the longer term? Do they have resources that can be dedicated to the effort? How do these stakeholder engagement objectives relate to the agencies’ other high priority work?

By way of example, an agency may decide that they wish to improve their comment analysis process or provide more content regarding the proposed action (e.g., links to maps from ArcGIS Online vs. simply pdfs).

Here is a sampling of recommendations categorized by near-term and mid to longer term windows. 

Tools for Discovering Content

Near-term recommendations

• Provide more consistent agencywide NEPA project website ‘look and feel,’ with more relevant content, conduct user surveys across multiple demographics
• Create/augment/update NEPA document repositories
• Enhance search and query functionality for NEPA repositories, including spatial search
• Make data/content easier to discover and access by strengthening internal data manage­ment/ data governance processes
• Make data/content download sites more visible; provide links from NEPA project sites to key agency download sites

Mid to Longer Term Recommendations

• Use analytics/AI to hone in on the most relevant content for NEPA project sites.
• Continue to augment/update NEPA document repositories, optimally offering multiagency capability.
• Enhance search and query functionality for NEPA repositories, using cognitive search and location intelligence.
• Continue to make data/content easier to discover and access across multiple disciplines by using analytics/AI.
• Continue efforts to make data/content downloads easier for stakeholders; expand use of APIs.

Tools for Analysis and Context

Near-term recommendations

• Augment existing NEPA project sites with data visualizations, including dashboards.Make the sites interactive.Use story maps (ESRI). Use animations.
• Use text analytics to streamline the agency comment analysis process.
• Improve existing processes for analyzing the NEPA effort, especially for developing the alternatives—collaborating more closely with stakeholders.
• Begin to incorporate more sentiment analysis and semantic analysis into the NEPA process in order to get deeper insights into stakeholder perspectives.

Mid to Longer Term Recommendations

• Expand the use of data visualization and location intelligence in highly interactive sites so that they become standard. Harvest the information from interactive sessions to inform the alternatives analysis.
• Use text analytics and associated advanced analytics/AI to streamline the agency comment analysis process across all channels.
• Continue to improve NEPA analysis, especially in developing the alternatives and collaborating more closely with stakeholders.
• Fully incorporate sentiment analysis and semantic analysis into the NEPA process.

Tools for Communication and Collaboration

Near-term recommendations

• Begin to use stakeholder engagement software. Identify the key requirements, scope of effort; consider some proofs-of-concept. Perhaps, at first, limit to NEPA and land use planning. Gather the key documents/databases etc. that contain stakeholder information.
• Augment/enhance existing processes for import and use of external data (e.g., from citizen science), including workflows that make data available to NEPA project staff.
• Consider development of a ‘master calendaring’ system across the land use planning and NEPA programs, for the benefit of external stakeholders.

Mid to Longer Term Recommendations

• Expand use of stakeholder engagement software so that interrelationships among stake­holders and shared issues can be identified and leveraged. If initial scope was limited, then expand so that it’s interdisciplinary.
• Further streamline processes for import and use of external data (e.g., from citizen sci­ence).
• Expand the ‘master calendaring’ system to multiple organizational levels and program areas for the benefit of external stakeholders.

Key Takeaways

Several key takeaways can be derived from the research described in previous sections:

• The four agencies vary in their ability to fulfill even partially the nine NEPA stakeholder engagement requirements; none currently provides adequate support for the highest stakeholder level: collaboration. If the agencies implement the recommendations in this section that best match their needs (and adapt/add to, as appropriate), they will move closer to providing that type of support.
• Achieving enhanced NEPA stakeholder collaboration requires improving overall agency service delivery (e.g., data and information), not simply modernizing the key apps that support NEPA.
• Policy guidance at the federal level, as well as department and agency guidance for the four resource management agencies, emphasizes stakeholder engagement and advocates the use of appropriate technology (here, analytics/AI) to solve mission needs.

As we have seen, there is no ‘magic’ path forward, nor are solutions simply of a technical nature. Analytics and AI will play an increasingly important role in the future, but in the context of stakeholder engagement they may be mostly ‘back office’ enablers, not necessarily visible to most stakeholders.

This report joins a library of IBM Center research focusing on how technology and analytics can improve decision-making, including: Silo Busting: The Challenges and Successes of Intergovernmental Data Sharing; More Then Meets AI Part I & Part II; Integrating Big Data and Thick Data to Improve Public Service Delivery;
and From Data to Decisions I & II.

We hope that the analysis and recommendations outlined in Jenna Yeager's timely report helps government agencies and stakeholders take advantage of evolving capabilities and enhance environmental decision-making using technology and analytics.