Introduction

The National Structure Inventory (NSI) provides a point-based inventory of structures used to assess the consequences of natural and technological hazards. While flood risk analysis remains the primary application, the dataset supports life safety and economic consequence modeling across a wide range of hazards.

The NSI is designed to provide a consistent baseline for consequence modeling. It performs best when used for regional or aggregate analyses and when supplemented with local data for higher-resolution applications. The NSI team recommends careful review of the base data and improvement of the data before use for local scale studies.

This document addresses common questions related to NSI attributes, limitations, and known issues. It also provides guidance for evaluating and refining NSI data for specific analyses. Users are encouraged to review the quality checks and known limitations sections when results appear unexpected.

Frequently Asked Questions

General Overview

What data is included in the NSI?

The NSI provides estimates of structure value, structure attributes, and the number of occupants associated with each structure. For a complete list of fields and definitions, refer to the technical documentation.

What is the coverage of the NSI?

The NSI includes the 50 states and the District of Columbia. Due to data availability and consistency constraints, U.S. Territories are not currently included.

What is the price level of the NSI?

NSI dollar values represent estimated replacement costs as of approximately August 2025.

NSI values reflect full and depreciated replacement cost, not market value. They exclude land value and broader market effects such as location premiums.

How reliable are the NSI estimates?

NSI is not intended for legal or regulatory determinations or for analyses requiring precise structure-level accuracy.

It is designed as a baseline inventory for screening-level and regional analyses. Users are expected to review, validate, and refine the data as needed for their specific application.

NSI estimates are generally more reliable at larger spatial scales, where regional assumptions and distributions better reflect aggregate conditions.

Structure and Spatial Questions

What is a "structure" in the NSI?

A structure is a modeled building location representing a physical structure such as a home, business, or public facility. Each structure includes attributes describing its size, use, value, and population.

Are NSI structures actual buildings?

NSI structures are modeled representations derived from available data sources, such as parcels, building footprints, Census data, and business datasets. While many structures closely align with real buildings, they should not be interpreted as authoritative building records.

Detached garages are typically not represented as separate structures. Their value is incorporated into the associated primary residence.

How accurate are structure locations?

Structure locations are based on building footprint data, parcel data, and other sources. In most cases, structures are located at the centroid of a building footprint. In some cases, structures may be placed at the centroid of a parcel when footprint data are unavailable.

All input datasets include some degree of error, such as missing or incorrectly identified footprints. Structures may also be located in areas that are detrimental for hazard modeling, such as piers or covered bridges.

Why are there sometimes multiple "structures" located at the same location?

Multiple structures may be assigned to the same footprint when different occupancy types are identified within a single building or parcel, such as residential units above ground-floor retail.

In these cases, square footage, value, and population are divided among the occupancy types to better represent mixed-use buildings. See the Known Limitations section for additional detail.

Population and Exposure Questions

What do NSI population fields represent?

NSI population fields represent a “most likely estimate” of the number of people present at a structure during typical daytime and nighttime conditions. Population is also categorized by age: under 65 and 65 and older.

These estimates are derived from Census data and other sources and are distributed across structures based on occupancy type and building characteristics. Refer to the technical documentation for additional detail.

Why might population seem too low or too high?

Differences between NSI estimates and expectations are typically driven by:

• Misclassification of occupancy type
• Inaccurate or incomplete building size estimates
• Limitations in available population data
• Data lag in rapidly growing or declining areas

Some areas, such as military installations, may lack reliable worker population data, which can affect estimates.

Does the NSI include estimates of transient population?

The NSI includes residential population estimates and worker population estimates, derived from LEHD data. Additional sources are used for facilities such as hospitals, prisons, schools, and hotels.

However, the NSI does not explicitly model shoppers, seasonal tourists in rental homes, or outdoor populations. Users may need to supplement NSI data for analyses where these populations have the potential to change risk characterization.

Valuation and Economic Use Questions

What does the structure value represent?

The NSI provides both full, non-depreciated, replacement value and depreciated replacement value. These values represent the estimated cost to rebuild the structure and may differ significantly from market value. NSI values exclude land value and do not account for market conditions such as demand or location desirability.

USACE analyses typically use depreciated replacement values, while FEMA HAZUS applications often use full replacement values.

How are structure values assigned?

Structure value is primarily estimated by multiplying building square footage by a cost per square foot. Cost assumptions vary based on occupancy type, building characteristics such as size and age, and regional economic factors such as labor costs and income levels.

Assessed values and real estate market prices are not used directly in these estimates. See the technical documentation for additional details.

Do values vary across regions?

Yes. Regional variation reflects differences in construction costs and economic conditions.

For example, single-family residential values are influenced by tract-level income, and all structures incorporate county-level adjustments based on construction labor costs.

Recommended User Quality Checks

Reviewing Outliers

Users should review structures with extreme values, such as unusually high population or structure value.

Check:

• Structure placement relative to hazard areas
• Exposure estimates
• Consequence estimates

Visualizing key attributes such as population, structure value, and number of stories in a GIS environment can help identify anomalies.

Reviewing Imagery

Compare the structure inventory to recent aerial or satellite imagery.

Confirm that:

• Structures are located in plausible positions
• High-priority areas are adequately represented
• No important developments appear to lack coverage

Conducting a Survey

For higher-detail studies, users may conduct surveys by sampling structures using imagery or field inspection.

This can help validate:

• Structure attributes such as foundation type and height
• Occupancy classifications
• Structure quality and depreciation

For smaller study areas, it may be practical to update most structures directly. For larger areas, sampled observations can be used to adjust attribute distributions.

The NSI Survey Tool is designed to support this process. Refer to the Survey Tool documentation for additional details.

Known Limitations

Stacked Structures

Commercial and mixed-use areas often contain multiple occupancy types within the same footprint. The NSI represents each occupancy type as a separate structure.

This approach allows:

• More accurate representation of mixed-use buildings
• Separate modeling of residential and non-residential populations

Structure value and population are not duplicated. The primary tradeoff is a modest increase in structure counts. At the national scale, this effect is relatively small; NSI 2026 includes approximately 134 million structures compared to roughly 130 million utilized building footprints. The effect is greater in non-residential areas where mixed uses are common.

Users requiring exact structure counts may merge structures using shared footprint identifiers or use the provided footprint ID to count unique values.

Mobile and Manufactured Homes

Mobile and manufactured homes present unique modeling challenges. While parcel data may identify some of these structures, mobile home parks are often represented as commercial parcels or large parcels with limited detail. The NSI supplements these data with external datasets identifying mobile home park locations, but incomplete or misaligned source data often remains.

As a result, the NSI may undercount or mischaracterize mobile homes in some areas. Users should review these areas carefully and make adjustments where necessary.

This example shows a mobile home park where building footprint data and parcel data do not fully capture the number and location of structures. Some homes are missing from the inventory, while others are misclassified as traditionally framed structures.

Large Parcels

The NSI typically assigns occupancy types based on parcel-level information and then associates those uses with building footprints. This process introduces uncertainty in large parcels containing many buildings, such as military bases and mixed-use developments.

Potential issues include:

• Incorrect assignment of use types
• Increased occurrence of stacked structures

Users should review these areas with a level of scrutiny appropriate to their analysis.

Large Buildings and Footprints

Large or complex buildings may be misrepresented due to limitations in footprint data.

Potential issues include:

• Partial building coverage leading to underestimated square footage
• Multiple attached buildings represented as a single footprint
• Concentration of value and population at a single point

Users should review structures with large modeled square footage to confirm reasonable representation.

This example illustrates a large industrial building where the footprint geometry does not cleanly represent distinct structures or internal divisions. As a result, square footage, structure value, and population may be inaccurately aggregated or distributed across the site.

Agricultural and Rural Areas

Rural areas often have lower-quality input data.

Common issues include:

• Missing building footprints in wooded or remote areas
• Parcel data that does not distinguish between residential and agricultural uses

The NSI attempts to align residential structure counts with Census data, but local inaccuracies may remain. Users should ensure that reasonable use types and exposure values are represented within their study areas, particularly in Farm House scenarios where the NSI struggles to distinguish primary residences from appurtenant structures.

High Growth and Declining Areas

The NSI 2026 integrates datasets with varying vintages, generally from 2020 to 2025.

As a result:

• Recently constructed buildings may be missing
• Demolished structures may still be included
• Population estimates may lag recent trends

Users should adjust structure counts and attributes in rapidly changing areas when accuracy is critical.

This example shows a recently constructed residential development that is not fully represented in the NSI due to the lag between real-world construction and the availability of input datasets. As a result, structures and associated population may be underrepresented in rapidly growing areas.

Inconsistent Structure-Level Data Availability

Data availability varies widely across jurisdictions. Some attributes, such as construction year, are commonly available at the structure level, while others, such as foundation type, may rely on regional assumptions more often.

Users should verify that key attributes are reasonable for their study area and supplement them with local data when necessary.