Small-area estimation of food insecurity
Household survey results are combined with satellite, census and market covariates to estimate food insecurity for districts the survey sampled too thinly, with an uncertainty for each estimate.
The Global Centre for Risk and Innovation (GCRI)
Spatial statistics turns sparse surveys, soil samples and market prices into maps with stated uncertainty. In food systems it estimates food insecurity for small areas, predicts soil properties between sampled profiles and tests whether price anomalies or malnutrition cluster in space.
GCRI · Established · service level R0
Small-area estimation is a set of statistical methods that combine sparse survey data with mapped covariates to estimate conditions, such as food insecurity, for districts the survey did not sample adequately. With digital soil mapping and cluster tests, spatial statistics turns sparse samples into maps whose uncertainty is reported.
Spatial statistics and geostatistics analyse data that vary across space, accounting for spatial dependence. They estimate values at unsampled locations and test for spatial patterns. In food and agriculture, trade, resources and supply chains and climate and hazards, spatial statistics produce continuous hazard and exposure surfaces with uncertainty. They reveal geographic clustering of risks and vulnerabilities across sectors. The canonical description is on the STEM Commons.
Household survey results are combined with satellite, census and market covariates to estimate food insecurity for districts the survey sampled too thinly, with an uncertainty for each estimate.
Soil properties measured at sampled profiles are predicted across the landscape from terrain, climate and spectral covariates by regression kriging or machine learning, with prediction intervals mapped.
Spatial autocorrelation and hotspot tests show whether high malnutrition or price anomalies cluster, before any cause is proposed.
Each term as this page uses it. Where a standard-setting body defines the term, the wording follows that body.
| Term | Definition |
|---|---|
| Small-area estimation | Estimating indicators for areas with few survey observations by borrowing strength from related areas and auxiliary data. |
| Digital soil mapping | Predicting soil properties across a landscape from sampled profiles and covariates such as terrain, climate and imagery. |
| Spatial autocorrelation | The tendency of nearby places to have similar values, which statistical tests must account for. |
Plains, deltas and basins, the corridors that carry food to ports, and how far each division sits from one, shown for West Africa as an example geography for this subject.
Data from the source register feeds the methods; calibration and an independent test gate the estimate and its limits; the decision stays with the competent authority.
Fig. 01Nodes carry their type, metrics and state; connectors show what feeds what; the gates are the validation steps; the timeline is one application. Design of the method's use, not a record of a run.
Datasets and records used by spatial statistics and geostatistics on this platform, with the role each plays and the label its values carry. Sources are listed in the source register.
| Data | Role | Label |
|---|---|---|
| Household and nutrition surveys | Response variables | OBSERVED |
| Soil profile databases | Point observations | OBSERVED |
| Satellite covariates | Predictors | MODELLED |
| Market price monitoring | Spatial price fields | OFFICIAL |
Predictions are tested using spatial cross-validation that respects dependence. Uncertainty maps are compared with independent observations.
| Check | How it is done |
|---|---|
| Spatial cross-validation | Blocks of locations are withheld in turn so spatial dependence does not inflate apparent skill |
| Uncertainty maps | Every published map has a companion map of its prediction error |
| Scale and zoning tests | Results are re-run at different aggregation units to check that patterns are not artefacts of boundaries |
Results depend on the scale and zoning of analysis. Sparse or clustered sampling limits prediction in unsampled areas.
Space and Earth Observation stewards spatial statistics and geostatistics for the Consortium and reviews how it is applied across platforms.
The current service level is R0: information and enquiries. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.