Calculate the total population affected by floods in Peru during February 2018

Calculate the total population affected by floods in Peru during February 2018

Task conventions: Use the frozen 2018 Peru population raster and supplied event flood_peru. Population source values are people per original cell. Keep its entire original grid; no population resampling. Sample the original flood raster at each original population-cell centre by containing pixel (nearest alignment); no max/bilinear interpolation or inferred flooded fractions. GFD bands are 1 flooded (0/1), 2 duration in days, 3 clear-view count, 4 clear-view fraction, 5 permanent water (0/1). Observe a cell only if population is finite/unmasked, bands 1,3,5 are finite/unmasked, clear views >0, and permanent water=0. Observed band1=1 is flooded; observed band1=0 is non-flooded. Everything else is unknown/excluded, never non-flooded or zero population. Calculate original full geographic-cell WGS84 ellipsoidal meridian/parallel area in km2; density is counts/area and counts are density*area as appropriate. Sum original/derived population over flooded cells. For the density comparison use group population divided by group physical area, not an unweighted mean of densities. This explicitly defines affected as people in observed flooded cell centres, not proven personal impacts. The event covers only part of the country/month: report the observed-footprint result, never an exhaustive national total. The supplied Bangladesh event's footprint includes northern Bangladesh despite its Tibetan event label; inspect actual coverage. Permanent water, cloud gaps and unobserved regions do not establish absence of flooding. No current source substitutions. Internal benchmark only.



Publish a full native-grid selected-population raster: people per cell for observed flooded cells, zero for observed non-flooded cells, NoData elsewhere. Also publish classification (1 flooded, 0 non-flooded, NoData unknown) on the same grid. End with fenced JSON {count:observed flooded cell count,unknown_count:all other unobserved/excluded population-grid cells,coverage_note:string,selection:{collectionId,itemId,assetKey},classification:{collectionId,itemId,assetKey},metrics:{selected_people:number,nonflood_people:number,selected_area_km2:number,nonflood_area_km2:number,selected_density:number,nonflood_density:number}}. Label classification unit as dimensionless and selected population as people per cell. Maps optional. Do not round machine-readable metrics to integers. Explicitly state incomplete country coverage and the 2018 source/event dates.

Delivery format: provide the answer object with all fields specified above as final_response on assess_result, alongside the inspected final artifact_refs. This replaces the JSON-code-block formatting requirement above. Blue attaches your explicit selection to the accepted final answer. Finish with a concise ordinary-language answer agreeing with that selected result; do not duplicate the JSON in chat.

Supplied immutable inputs:
[
  {
    "name": "Peru population count, WorldPop 2018",
    "edition": "2018 UN-adjusted population counts, aggregated to 1 km cells",
    "units": {
      "band1": "people per cell"
    },
    "collectionId": "blue-generated--[resource id withheld]",
    "itemId": "[resource id withheld]",
    "assetKey": "data"
  },
  {
    "name": "GFD observed flood 4569, 2018-02-01 to 2018-02-21",
    "edition": "Original five-band GFD v1 benchmark snapshot, not a current flood map; benchmark-only use",
    "units": {
      "band1": "binary flooded flag",
      "band2": "days",
      "band3": "observations",
      "band4": "fraction",
      "band5": "binary permanent-water flag"
    },
    "collectionId": "blue-generated--[resource id withheld]",
    "itemId": "[resource id withheld]",
    "assetKey": "data"
  }
]

{
  "expectedCount": 113,
  "actualCount": 113,
  "expectedMetrics": {
    "selected_people": 2659.5894656293094,
    "nonflood_people": 826501.1746531725,
    "selected_area_km2": 92.91052422702717,
    "nonflood_area_km2": 16624.819734452987,
    "selected_density": 28.62527671386929,
    "nonflood_density": 49.71489543073636
  },
  "actualMetrics": {
    "nonflood_people": 826501.1746531725,
    "selected_people": 2659.5894656293094,
    "nonflood_density": 49.71489541397814,
    "selected_density": 28.6252766519842,
    "nonflood_area_km2": 16624.81974005699,
    "selected_area_km2": 92.91052442789078
  },
  "computationCorrect": true,
  "answerFulfilled": true,
  "adjudication": null,
  "qualification": null,
  "reasons": []
}

Does the native-grid method preserve the population grid, classify observed flooded versus non-flooded cells correctly with unknowns left out, and produce additive population and area summaries on a representative Peru overlap area?

The trial preserved the full original Peru population grid at 1522 by 2199 cells with 0.0083333333 degree resolution and aligned the flood, clear-view, and permanent-water bands to that grid with nearest resampling. Outside the observed event footprint the aligned flood bands remained missing. The intermediate class_code used -1 as an internal sentinel for unknown or excluded cells, and the reclassify step converted those cells to NoData in the final classification output. The final classification produced 20,296 observed cells, with 113 flooded cells and 20,183 observed non-flooded cells; 3,326,582 population-grid cells remained unknown or excluded. The selected-population raster retained people per cell only for flooded cells, wrote zero for observed non-flooded cells, and left unknown cells as NoData. Exact trial summaries provide additive totals for observed population and observed area, so non-flooded population and area can be derived by subtraction without changing the method.

The published native-grid outputs report 113 observed flooded population-grid cells in Peru for the supplied GFD observed flood 4569 event dated 2018-02-01 to 2018-02-21. Those flooded cells contain 2,659.5894656293094 people over 92.91052442789078 km2, for a flooded-group density of 28.6252766519842 people per km2. The observed non-flooded cells total 20,183 by subtraction from 20,296 observed valid cells, and they contain 826,501.1746531725 people over 16,624.81974005699 km2, for a non-flooded observed density of 49.71489541397814 people per km2. The remaining 3,326,582 cells on the full Peru population grid are unknown or excluded. Sources: Peru population count, WorldPop 2018 benchmark fixture (Blue-hosted WorldPop/CIESIN benchmark input; source revision not provided; retrieved 2026-10-07) and GFD observed flood 4569 benchmark fixture, 2018-02-01 to 2018-02-21 (Blue-hosted Global Flood Database/Cloud to Street benchmark input; source revision not provided; retrieved 2026-10-07).

The published native-grid outputs report 113 observed flooded population-grid cells in Peru for the supplied GFD observed flood 4569 event dated 2018-02-01 to 2018-02-21. Those flooded cells contain 2,659.5894656293094 people over 92.91052442789078 km2, for a flooded-group density of 28.6252766519842 people per km2. The observed non-flooded cells total 20,183 by subtraction from 20,296 observed valid cells, and they contain 826,501.1746531725 people over 16,624.81974005699 km2, for a non-flooded observed density of 49.71489541397814 people per km2. The remaining 3,326,582 cells on the full Peru population grid are unknown or excluded. Sources: Peru population count, WorldPop 2018 benchmark fixture (Blue-hosted WorldPop/CIESIN benchmark input; source revision not provided; retrieved 2026-10-07) and GFD observed flood 4569 benchmark fixture, 2018-02-01 to 2018-02-21 (Blue-hosted Global Flood Database/Cloud to Street benchmark input; source revision not provided; retrieved 2026-10-07).

The published native-grid outputs report 113 observed flooded population-grid cells in Peru for the supplied GFD observed flood 4569 event dated 2018-02-01 to 2018-02-21. Those flooded cells contain 2,659.5894656293094 people over 92.91052442789078 km2, for a flooded-group density of 28.6252766519842 people per km2. The observed non-flooded cells total 20,183 by subtraction from 20,296 observed valid cells, and they contain 826,501.1746531725 people over 16,624.81974005699 km2, for a non-flooded observed density of 49.71489541397814 people per km2. The remaining 3,326,582 cells on the full Peru population grid are unknown or excluded. Sources: Peru population count, WorldPop 2018 benchmark fixture (Blue-hosted WorldPop/CIESIN benchmark input; source revision not provided; retrieved 2026-10-07) and GFD observed flood 4569 benchmark fixture, 2018-02-01 to 2018-02-21 (Blue-hosted Global Flood Database/Cloud to Street benchmark input; source revision not provided; retrieved 2026-10-07).

The published selected-population raster is the required native-grid output over Peru. It preserves the frozen 2018 population grid and publishes people per cell for observed flooded cells, zero for observed non-flooded cells, and NoData elsewhere.

The published classification raster is the required native-grid 1/0/NoData output over Peru. It preserves the original population grid and records observed flooded cells as 1, observed non-flooded cells as 0, and all unknown or excluded cells as NoData.