Explore the same location across dry season, wet season, historical years, and modeled wetland-fill scenarios. The goal is to show the relationship between the land, water, and surrounding terrain — not to invent parcel-level facts.
Modeled environmental viewObserved, modeled, and scenario layers are clearly separated.
Surface water0% · River response0% · Groundfirm
Observed / mappedModeledScenario
0%
Standing water
0%
Water persistence
0%
Erosion risk
—
Composite risk
Watch the same place change across seasons and years.
Move the year and month controls, hover charts, or compare seasons. The graphics are linked to the same model driving the animation.
Seasonal rainfall pattern—
Modeled rainfallSelected month
Where the exposure comes fromCurrent state
Water through the year—
1935198520262035
Season comparisonRelative exposure
Scenario impactWhat changes when land is filled or blocked?
Dry season—
How the place can appear when moisture is low.
Wet season—
Modeled surface-water pressure after seasonal rain returns.
Filled/blocked—
Counterfactual scenario; not evidence that filling occurred.
What's driving the composite risk
Rainfall load
0%
Drainage capacity
0%
Dry-soil runoff
0%
How this demo model works: your location's region is matched to one of Ghana's four real climate zones (Northern Savanna, Transition/Forest-Savanna, High Forest, or Coastal — shown in the Regional climate context panel), each with its own documented annual rainfall range, harmattan severity, and drainage/flooding pattern. Once you start the simulation, rainfall and harmattan drift live toward that zone's actual wet- or dry-season character, not a generic one-size-fits-all Ghana pattern. Drainage stays under your control, since it's a property of the plot's infrastructure, not the weather.
Why the clock starts at 1935: that's not arbitrary — it's the earliest year in this project's own historical flood dataset, i.e. the real start of the record this tool has behind it. The clock counts years forward from there. What's not real: which specific year gets which weather. The season pattern and each zone's annual rainfall range are sourced; year-to-year variation within that pattern is an illustrative random walk, not a reconstruction of what actually happened in, say, 1962. Once the count passes the real present day, it says so plainly ("beyond recorded history") rather than quietly pretending to still be historical.
What's real here and what's derived: each zone's annual rainfall range and its harmattan/drainage description are sourced from regional Ghana climatology, not invented. The annual/seasonal rainfall scenario the simulation uses are a transparent derivation from those real annual ranges (concentrating most of the year's rain into the documented wet-season pattern) — a reasonable estimate, not a directly measured daily figure. I don't have per-town or per-coordinate historical weather records, so this is zone-level climatology, not this specific parcel's own measured history.
About "Use my GPS": the town/farmland label is a best-effort read of OpenStreetMap's place data for your coordinates (city/town/suburb tags vs. village/hamlet vs. neither) — it's a reasonable guide, not a verified land-use classification. Note that "rural" doesn't mean "lower risk" everywhere: Savanna farmland, for instance, has real documented flash-flood risk from low-infiltration clay soils despite being far from any town.
Going further: true per-parcel precision would mean pulling actual elevation, slope, and rainfall at this exact coordinate from DEM and satellite rainfall data (the kind of pipeline that extracts these from GeoTIFF rasters) rather than zone-level averages. That needs a backend with real elevation/rainfall raster data behind it, which isn't connected here — if you have that running, the drainage/erosion model could use this parcel's real slope and elevation instead of a slider.