From Landslide Maps to Landslide Intelligence

A landslide is rarely just a geological event. It can quickly become:

• A transportation problem when a road closes.
• A utility problem when water, sewer, power or communications infrastructure is damaged.
• An emergency-management problem when access or evacuation routes are compromised.
• A planning problem when development occurs in vulnerable terrain.
• A capital-planning problem when mitigation competes with other infrastructure investments.
• And ultimately a public-safety problem when people and property are exposed.

Landslide susceptibility is influenced by a combination of geology, slope, rainfall, soil saturation and seismic conditions, along with human activities including land-use change and infrastructure development. That makes landslide risk inherently cross-departmental.

The underlying research shows how rapidly this field is changing. Scientific activity around geospatial landslide susceptibility increased substantially during the last decade, particularly after 2019. Behind that growth is the convergence of several technologies that previously operated much more independently.

Diagram 1 — From Map to Intelligence

AI does not replace GIS. It expands what GIS can help government understand.

Local government ultimately cares about consequences. A highly susceptible slope in an isolated area may require monitoring. A similar slope immediately above a major highway, water-treatment facility or residential neighborhood has an entirely different significance.

That is where GIS becomes especially powerful. It allows agencies to combine the physical hazard with what sits above, below and around it. A GIS-based risk environment can potentially answer questions such as:

• Is there a road below the slope?
• Is that road an evacuation route?
• Is a major water line nearby?
• Are homes potentially exposed?
• Is development planned in the area?
• Has movement occurred previously?
• Are drainage problems documented?
• Is the area recently affected by wildfire?
• Are there critical communications assets nearby?
• Would failure isolate a community?

Diagram 2 — The Local-Government Risk Chain

The progression from hazard to consequence is where landslide mapping becomes a local-government management tool.

GIS becomes the common spatial language connecting all of those questions.

One of the most important lessons for government is that sophisticated technology should not be the starting point. A jurisdiction with incomplete landslide inventories, fragmented asset information and inconsistent terrain data may gain far more by improving its GIS foundation than by immediately deploying an advanced AI model. A more practical progression is incremental.

An organization cannot become meaningfully predictive until it first becomes connected.

This maturity model is particularly useful for local-government leaders because it prevents AI from becoming a technology project disconnected from operational needs. The objective is not to reach Level 6 as quickly as possible. The objective is to move to the level that produces better decisions for the organization.

Machine learning and deep learning are becoming increasingly important in landslide research. But government faces a different standard from an academic research project. A model may ultimately influence infrastructure spending, inspection priorities, development review, emergency planning, mitigation projects and communications with the public.

Government therefore needs to understand more than the final prediction. It needs to understand how that prediction was created.

Human expertise remains essential. Remote sensing and AI do not eliminate the need for geologists, geotechnical engineers, surveyors, GIS professionals, planners, public works personnel, emergency managers or field inspection. Instead, they can help those specialists determine where their limited time and resources may have the greatest value.

Treat AI-generated susceptibility as decision support rather than certainty. Before operational use, establish:

• Responsible department
• Authoritative datasets
• Data-update schedules
• Model documentation
• Validation requirements
• Confidence thresholds
• Escalation procedures
• Human review

The best system is not necessarily the one with the most sophisticated algorithm. It is the one staff trust enough to use.

The most advanced Geospatial-AI capabilities may still be evolving. Local governments do not need to wait.

Before making the next technology investment, ask:

• Do we know our highest-consequence slopes?
• Can we connect landslide susceptibility with roads, utilities and critical facilities?
• Do planning, public works, engineering and emergency management share the same risk picture?
• Do we know which datasets are authoritative?
• Where would updated remote sensing actually change a decision?
• Do we have enough historical information to evaluate predictive models?
• Could technology reduce unnecessary field inspection?
• Can any AI-generated recommendation be explained?
• Who acts when elevated risk is identified?
• And finally: what happens after the map identifies the problem?

If there is no clear answer, another map or another model probably isn't the immediate priority. A better workflow is.

The global research landscape points in a clear direction. Landslide susceptibility analysis is evolving from relatively static GIS mapping toward an increasingly connected environment incorporating GIS, Earth observation, remote sensing, cloud computing, big geospatial datasets and artificial intelligence.

But for local government, the most important transformation is not the technology itself. It is what that technology allows the organization to do.

That changes the role of GIS. GIS becomes less about producing the final map and more about connecting the information required to make a decision. The future of landslide management therefore will not simply be defined by better susceptibility maps. It will be defined by how effectively local governments turn spatial intelligence into better inspections, better infrastructure investments, better land-use decisions, better mitigation and better preparedness.