SpatialKat version 7.0 includes model updates for Canada Inland Flood, US Storm Surge, and introduces a new model, US Wildfire. All other existing Inland Flood, Storm Surge, and Tropical Cyclone models are unchanged. 

In addition to the new and updated models, software functionality updates have been included to improve performance, stability, configuration validation, and cross-platform (Windows and Linux) support.  

New Model: 

• US Wildfire 

New Software Functionality: 

• Performance improvements for large portfolios (server-side table processing, faster login, readable large-number formatting) 

• Startup configuration validation and clearer error reporting 

• Improved stability and resource management for large uploads and long sessions 

• Native Linux build and runtime support alongside Windows, with service-managed restarts 

• API enhancements, including wildfire and smoke parameters 

• Profile and exposure input handling improvements 

Model and Data Updates: 

• Updated Canada Inland Flood maps, scores, and defense layer 

• Updated US Storm Surge terrain data, maps, and scores 

• New historical events (2025 and 2026) 

• Updated 2025-OBS climate state 

Requirements 

Upgrading to SpatialKat version 7.0 requires reimport of all exposures.  

Updated Models 

Canada Inland Flood 

The Canada Inland Flood model update includes new flood maps for all return periods (10, 20, 50, 100, 200, and 500) across Canada based on new terrain data, now at 10m resolution. Spatially varied defense data is also included in the model, replacing the previous uniform defense assumption. Scores and demand surge have been updated. Four historical events have also been updated in an effort to ensure the combination of the new events and new hazard maps reflect the observed impacts of the events well. 

US Storm Surge 

The US Storm Surge update includes new terrain data as well as flood maps for all return periods (10, 20, 50, 100, 200, and 500). Scores have also been updated. The storm surge model methodology is unchanged; this is an update to the underlying terrain (elevation) data.  

US Wildfire 

SpatialKat version 7.0 introduces the KatRisk contiguous US (CONUS) wildfire model, a high-resolution model that includes an explicit urban conflagration component and a comprehensive vulnerability model. The fire-spread model produces physically realistic footprints, including in hard-to-model urban areas, drawing on proprietary fuel layers, high-resolution meteorological data, and the Technosylva fire-spread model. Simulated fire frequency and size are calibrated to be representative of the historical record. 

A distinguishing feature of the model is that the spatial correlation of ignition points is driven by the same KatRisk stochastic precipitation dataset that drives the KatRisk flood model, creating an explicit correlation between flood and wildfire risk. Vulnerability curves are built using a copula method that tracks the correlation between building characteristics and two hazard inputs, rate of spread and conditional flame length, capturing radiant heat, direct flame contact, and ember-transport damage pathways. The model was validated using a three-tier approach comparing historical hazard, structural loss, and financial loss targets to modeled results. 

New Software Functionality 

Performance for large portfolios 

Table performance for large portfolios has been significantly improved through server-side processing, with less exposure and financial data transferred to the browser and financial child rows loaded only when an exposure is expanded. Non-visible modules are now deferred until their tabs are opened, reducing login time. Large AAL, standard deviation, OEP, and AEP values now display as readable formatted numbers rather than scientific notation. 

Configuration validation and error reporting 

SpatialKat now performs comprehensive startup validation of required configuration paths, realm data files, and CSV schemas. A blocking configuration screen identifies any missing item, its configuration key, the expected location, and corrective guidance, and configuration can be reloaded and rechecked without restarting the server. Error messages for configuration, model-availability, licensing, preprocessing, and background-job failures have been made clearer and more specific. 

Stability and resource management 

Memory stability during very large exposure uploads and long-running sessions has been improved, with cleanup of temporary files and isolated temporary directories for the web and API components. Large downloads are now disk-backed to reduce memory pressure, and the web server and API daemon are restarted together, with support for service-managed restarts via Windows Task Scheduler and Linux systemd. 

Native Linux support 

SpatialKat version 7.0 adds native Linux build and runtime support across engines and wrapper scripts, alongside continued Windows support, including the SpatialKat and Underwriting APIs and their shared components. Build and version metadata are embedded in Linux executables for operational traceability. 

API enhancements 

The API has been extended to support the new wildfire and smoke parameters and workflows. Combine and Marginal Impact operations can now use analysis IDs found in the job queue rather than only analyses already registered in the completed-analysis list, and API error reporting for configuration, model availability, licensing, and job failures has been improved. 

Profile and exposure input handling 

The profile master list has been migrated from profile.csv to profile.json (with compatibility retained for existing deployments) and profile-list caching has been added. Handling of special characters, accented text, punctuation, CSV delimiters, and HTML-sensitive characters in user and exposure inputs has been improved, and SQL Server output processing now uses configurable transaction batching with consistent parameter handling. 

Model and Data Updates 

Historical event updates 

Historical events have been updated across regions: US Inland Flood events for 2025 and 2026; Australia Inland Flood events for 2025; and a Central America and a Mexico (CENMEX) Inland Flood event for 2025. 

2025-OBS climate state 

The 2025-OBS climate state has been updated, incorporating additional observations gathered since the climate state was first released. 

YEAR_BUILT input range extended 

The valid range for the YEAR_BUILT exposure input has been extended to accept values up to 2035. In earlier versions, a YEAR_BUILT value above 2025 fell outside the valid range and defaulted to unknown. YEAR_BUILT affects vulnerability mapping for wind-related perils only (AFTC, ASTC, NATC, NITC, OCTC, NASW, and NATO).