The Hydrogeological Risk (Flood Risk) KPI estimates the flood risk for a site based on the historical record of large flood events from the Dartmouth Flood Observatory (DFO) Global Active Archive (1985 -- present), weighted by severity and intersected with HydroBASINS Level 12 watershed boundaries. The result is a risk index from 0 to 20, classified into 5 discrete risk levels.
Floods are among the most devastating natural hazards globally, with growing frequency and intensity driven by climate change -- increasing extreme precipitation events, altered hydrological regimes, and expanded impermeable surfaces. This KPI integrates historical data with future projections from a Machine Learning model trained on CMIP6 climate variables and SSP scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5), providing both current risk classification and probabilistic future outlooks.
The quality scale maps risk levels as follows:
| Level | Flood Risk Index | Risk Level | Interpretation |
|---|---|---|---|
| A (Excellent) | 0 -- 1 | 1 | Minimal flood risk -- no significant events on record |
| B (Good) | 1 -- 5 | 2 | Low risk -- rare or low-severity historical events |
| C (Moderate) | 5 -- 8 | 3 | Moderate risk -- occasional events in the watershed |
| D (Poor) | 8 -- 15 | 4 | High risk -- multiple significant events recorded |
| E (Critical) | 15 -- 20 | 5 | Extreme flood risk -- frequent and severe events |
Sites with risk level >= 3 are flagged as at-risk by the platform threshold.
The flood risk index is computed by intersecting the site polygon with HydroBASINS Level 12 watershed boundaries, extracting all flood events from the DFO Global Active Archive (1985 -- present) that overlap the watershed, and weighting each event by severity (1 = limited impact, 2 = moderate, 3 = severe). The weighted sum is then normalized to a 0--20 scale and classified into 5 discrete risk levels.
For future projections, a Machine Learning model trained on historical flood events and CMIP6 climate variables computes per-year event probabilities under three SSP scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5). The combined 2-year probability is calculated as:
where and are the annual probabilities for two consecutive years, de-duplicated to avoid double-counting overlapping risk.
Data are sourced from the DFO Global Active Archive (event-level polygons, 1985 -- present), HydroBASINS Level 12 watershed boundaries (HydroSHEDS/WWF), and CDS CMIP6 climate projections (~100 km resolution, 2026--2100).
Gauge. A semicircular gauge showing the flood risk index for the site (ROI) and its control area (CA).
Purpose: What is the historical flood risk level at this site, and how does it compare to the surrounding control area?
Description: The gauge displays two arcs on a 0--20 scale segmented into 5 quality zones (A--E). A circular marker indicates the site's flood risk index value. Below the gauge, text shows the risk index and quality score. A colored A--E badge indicates the quality grade. The left panel also shows the past-events count as a separate badge. Topic and regulatory reference labels appear in the card header.
9.4Flood Risk Index
How it's calculated: The displayed value is the weighted sum of DFO flood events in the watershed, normalized to [0--20]. The quality grade is derived from the 1--5 risk level classification mapped to A--E.
Note: This indicator is inverted -- lower values indicate safer conditions.
Legend:
| Level | Flood Risk Index | Color | Meaning |
|---|---|---|---|
| A | 0 -- 1 | ■ #00A67A | Minimal risk |
| B | 1 -- 5 | ■ #00DF80 | Low risk |
| C | 5 -- 8 | ■ #FFD21E | Moderate risk |
| D | 8 -- 15 | ■ #FF8B16 | High risk |
| E | 15 -- 20 | ■ #FF367F | Extreme risk |
Interpretation example:
If the gauge shows a site index of 9.4 (grade D) and a control area index of 3.1 (grade B), the site lies in a watershed with significantly more historical flood activity than its surroundings -- indicating elevated infrastructure and ecosystem risk that warrants monitoring of drainage systems and vegetation cover.
Slider Card. A panel adjacent to the gauge showing ML-modeled event probabilities for the next 2 years (or a selected future year) under a chosen SSP scenario.
Purpose: How likely is a flood event at this site in the near future under different climate pathways?
Description: Two dropdown selectors allow the user to choose between "Next 2 years" (combined 2-year probability) or a specific future year, and to select an SSP scenario (SSP1-2.6, SSP2-4.5, SSP5-8.5). Two probability values are displayed: the probability of any event, and the probability of a severe event. Values are shown as percentages. This panel is only visible when the site has the future projections functionality enabled.
How it's calculated: Probabilities are generated by the ML model from CDS-CMIP6 future climate variables. The 2-year combined probability uses de-duplicated union: P(2y) = P1 + P2 - P1 x P2. Severe event probabilities are capped at 70% of the corresponding "all events" probability.
Interpretation example:
If SSP5-8.5 shows 42.3% probability of any event and 18.7% probability of a severe event within 2 years, the site faces a non-negligible near-term flood risk under high-emission pathways -- the gap between scenarios is useful for quantifying climate mitigation benefits.
Bubble Chart. A scatter/bubble visualization showing historical flood events from 1985 to present, grouped by year and event category.
Purpose: When did flood events occur, how many per year, and how severe were they?
Description: Each bubble represents a group of events in a given year sharing the same cause category (rain, dam break, avalanche, ice melt, snow melt, tide, tsunami, other). The X-axis shows the year, the Y-axis shows the number of events. Bubble radius is proportional to the total severity of events in that group. A skull icon is overlaid on bubbles where deaths were recorded. Hovering shows the number of deaths in a tooltip. The chart can be expanded to full-screen. Tab controls switch between Chart, Summary, and Table views.
How it's calculated: Events are grouped by year and main cause category. Bubble radius scales with total severity relative to event count. Colors are assigned by category.
Interpretation example:
If a large bubble with a skull icon appears at year 2002 with Y = 3 events, this indicates 3 flood events in 2002 of the same category with significant severity and casualties -- a historical hotspot requiring attention in site risk planning.
Summary Dashboard. A grid of metric cards and distribution tables summarizing the statistical profile of historical flood events at the site.
Purpose: What is the aggregate risk profile -- how many events, victims, average severity, and dominant cause?
Description: The top grid shows 4 cards: Risk Level (X/5), Risk Index (numeric), Total Events, and Total Victims. Below, two more cards show average victims per event and average severity. Two text rows show the most common cause category and the most recent event date. Two sub-tables show decade distribution (events, victims, % by decade) and severity distribution (count and % by severity class).
How it's calculated: All values are computed from the full list of historical DFO flood events matched to the site's watershed. Averages, totals, and distributions are derived from event-level data on casualties, severity, cause, and dates.
Interpretation example:
If the summary shows 12 total events, 3 high-severity events, and "Rain" as the most common cause, the site's watershed is exposed to recurrent precipitation-driven flooding -- nature-based solutions targeting water retention and slope stabilization are most relevant.
Data Table. A paginated table listing all historical flood events matched to the site's watershed, with full event details.
Purpose: What are the specific historical events -- dates, causes, severity, and casualties?
Description: Columns include: Event Name, Started (date), Ended (date), Deaths, Cause, Category, Severity (1--3). Rows can be copied. A CSV download button exports the full table. This tab is accessible via the "Table" switch in the events chart area.
How it's calculated: Data comes directly from the DFO archive matched to the site's HydroBASINS Level 12 watershed.
Interpretation example:
If the table shows an event in March 2010 with severity 3, 24 deaths, cause "Rain", the site's watershed experienced a catastrophic precipitation-driven flood in 2010 -- this should be cross-referenced against local infrastructure investments and flood mitigation measures made since then.
Assessment Sidebar Row. A quality grade row in the Assessment sidebar panel, summarizing the flood risk quality grade for the site.
Purpose: Provides an at-a-glance A--E quality grade for flood risk within the overall site assessment summary.
Description: The row is labeled "Flood Risk" and shows an A--E colored badge derived from the site's flood risk index. Clicking the row opens the detailed Risks & Dependencies section.
How it's calculated: The grade is derived from the site's flood risk index mapped to the 5 quality levels using the inverted scale.
Note: This indicator is inverted -- lower values indicate safer conditions.
Interpretation example:
If the sidebar shows a "D" grade for Flood Risk, the site's watershed has a flood risk index between 8 and 15, indicating high historical flood exposure requiring active risk management strategies.
Highlights Table Row. A numeric comparison row in the KPI comparison table within the Highlights section, showing flood risk values for site and control side by side.
Purpose: Allows direct numeric comparison of flood risk between site and control area with quality grading.
Description: The row is labeled "Flood Risk" and falls under the E3-E4 topic group. It shows the site's risk level value, the control area risk level value, and a colored A--E grade badge. Values use the inverted quality scale.
How it's calculated: Uses the site and control area risk levels (1--5 scale) mapped through the inverted quality ranges to produce the A--E grade.
Note: This indicator is inverted -- lower values indicate safer conditions.
Legend:
| Level | Risk Level | Color | Meaning |
|---|---|---|---|
| A | 1 | ■ #00A67A | Very low risk |
| B | 2 | ■ #00DF80 | Low risk |
| C | 3 | ■ #FFD21E | Moderate risk |
| D | 4 | ■ #FF8B16 | High risk |
| E | 5 | ■ #FF367F | Extreme risk |
Interpretation example:
If the row shows site risk level 3 (grade C) vs control 1 (grade A), the site is in a flood-prone watershed while the control area has minimal historical flood exposure -- a meaningful divergence that may reflect proximity to a river or low-lying terrain.
Highlights Card. A compact summary card in the Highlights section showing the hydrogeological risk quality icon for the site.
Purpose: Provides an instant visual signal for hydrogeological risk quality within the broader ESG Highlights overview.
Description: The card is labeled "Hydrogeological" and shows a quality icon (1--5 scale) for the site. The topic label reads "E3-E4".
How it's calculated: Derived from the site's risk level value mapped to the 1--5 icon scale using the inverted quality ranges.
Interpretation example:
If the Hydrogeological card shows 4 icons highlighted (out of 5), this indicates a high risk level -- the site's watershed has significant historical flood exposure that is elevated compared to the A--E quality scale.
Map Layer. An interactive raster overlay showing the discrete flood risk level across the site and surrounding area, using 5 colors corresponding to risk levels 1--5.
Purpose: What is the spatial distribution of flood risk levels across the site territory?
Description: The map displays a colored overlay where each pixel represents the flood risk level of the HydroBASINS Level 12 watershed containing that location. Because flood risk is computed at the watershed scale, the entire watershed appears in a uniform color. A tooltip explains the layer represents flood risk based on historical flood events. Below the map, a legend lists the 5 risk levels with their colors. The map layer is selectable from the Risk sublayer group.
How it's calculated: The risk level (1--5) is computed from the DFO event archive weighted sum, then colored using a discrete colormap. The image is pre-computed and served as a raster overlay.
Legend:
| Level | Color | Label |
|---|---|---|
| 1 | ■ #7F9127 | Very Low |
| 2 | ■ #147B20 | Low |
| 3 | ■ #F7A333 | Moderate |
| 4 | ■ #F85621 | High |
| 5 | ■ #FB1C19 | Very High |
Interpretation example:
If the map shows the entire site watershed in orange (level 3 -- Moderate), the watershed has historical flood activity placing it at moderate risk. Nearby watersheds in green indicate lower-risk areas -- useful context for portfolio-level site comparison.
Slider Card. A horizontally scrollable card panel in the Map Detail section, appearing beneath the map when the Flood Risk layer is active, showing individual risk zone cards for each identified flood-prone area within the site.
Purpose: Which specific areas within the site are in flood-prone zones, and what are their risk levels?
Description: Each card shows a color swatch (risk level color), the risk level label as title, and a description line with the risk index, level, and score. Below, the area in hectares is shown. If there are no flood-prone areas, an empty-state message is displayed. A summary line appears in the panel header when data is available. Cards scroll horizontally with smooth snap behavior.
How it's calculated: Each card represents a distinct land polygon intersected with a flood risk zone. The index value is the site-level flood risk per area; the risk level and A--E score badge are derived from the standard quality mapping.
Interpretation example:
If three cards appear showing levels 2, 3, and 4 across different parts of the site, the site territory spans multiple risk zones -- the level 4 area (High risk) requires priority attention for drainage design and vegetation buffer planning.
Highlights Card. A compact card in the iLoveNatura Widget sidebar showing flood risk level and event probability for a selected site.
Purpose: What is the flood risk at this site, and what is the probability of a flood event?
Description: The card shows a wave icon colored by risk quality, the title "Flood Risk" and subtitle "Flood Risk Assessment". Below: a 2-column mini-grid shows Risk Level (numeric) and Event Probability (%). Up to 3 recent events are shown as tags with their main cause label. A map-layer icon indicates clicking the card activates the flood risk map layer.
How it's calculated: Risk level and event probability are derived from the site KPI data. Color is derived from the quality ranges with the inverted flag. The quality label (A--E) follows the standard mapping.
Note: This indicator is inverted -- lower values indicate safer conditions.
Interpretation example:
If the card shows Risk Level 2.0 (grade B) and Event Probability 12%, the site has low historical flood risk with a low near-term probability -- a favorable signal for site safety and ecosystem stability.
| Source | Provider | Coverage | Resolution | Period |
|---|---|---|---|---|
| Global Active Archive of Large Flood Events (DFO) | Dartmouth Flood Observatory, University of Colorado | Global | Event-level polygons | 1985 -- present |
| HydroBASINS Level 12 | HydroSHEDS / WWF | Global | Watershed polygons | Reference: 2013 |
| Climate Data Store CMIP6 | Copernicus Climate Change Service (CDS) | Global | ~100 km (model-dependent) | Projections 2026 -- 2100 |
The site polygon is intersected with HydroBASINS Level 12 boundaries to identify the contributing watershed(s). All flood events recorded in the DFO Global Active Archive from 1985 to present that overlap the watershed are extracted. Each event is assigned a severity weight (1 = limited impact, 2 = moderate, 3 = severe, based on estimated damage to structures and population). The weighted sum of events is normalized to the [0, 20] range to produce the flood risk index. The index is then classified into 5 discrete risk levels (1 = Very Low through 5 = Very High).
For future projections, a Machine Learning model trained on historical flood event data and meteorological/climate variables computes the probability of a flood event occurring in the coming years for each SSP scenario. Future meteorological inputs come from the CDS CMIP6 dataset. The model produces per-year event probabilities (%) for "all events" and "severe events" per scenario. Severe event probabilities are capped at 70% of the corresponding "all events" probability to prevent systematic overestimation of tail risks.
For the map layer, the risk level (1--5) is computed from the DFO event archive weighted sum and colored using a discrete 5-color colormap. Because flood risk is computed at the watershed scale, the entire watershed appears in a uniform color. The image is pre-computed and served as a raster overlay.