Bhutan: geospatial and demographic evidence for impact assessment

Bhutan's census, household surveys and spatial records provide a useful basis for environmental and social appraisal. Their analytical value comes from connecting a development activity with the people, livelihoods, services and ecosystems affected through a specified pathway. Relevant study areas include project footprints, downstream catchments, transport networks and seasonal resource-use areas.

This discussion identifies sources and study-design considerations for hydropower, roads, conservation and urban development. The flood example draws on a published analysis; the other applications describe research designs. Source specifications refer to provider records checked in October 2026. Record the selected edition, access conditions and study-specific suitability in the dataset register.

The baseline assessment method explains how to select evidence for appraisal. The questionnaire guide and sampling guide develop the primary research needed to connect national and spatial context with affected people's circumstances.

Select evidence for the decision

The agriculture and livestock catalogue adds livelihood and production context. The 2022 Gross National Happiness survey report provides a complementary wellbeing framework spanning nine domains and 33 indicators. Select outcomes that correspond to the intervention, such as time use, service access and community relationships.

Reconcile statistical and spatial geography

The poverty-mapping report combines 269 original gewog and town units into 197 reporting areas. Annex 1 supplies the aggregation crosswalk. Reconstruct those combined areas before joining estimates to polygons, and retain the reported uncertainty. An area's estimated poverty rate supports geographic prioritisation; household eligibility requires household evidence. NSB and World Bank, 2023

Survey coverage and statistical support also differ. Q3 2026 labour-force fieldwork covered all 20 dzongkhags, while NSB publishes headline estimates nationally because the sample has limited precision at finer levels. Use published reporting domains when describing local economic conditions. NSB Q3 report

Maintain a crosswalk with official code, source name, alternative spelling, unit type, parent unit and geometry vintage. A gewog, thromde and dzongkhag describe different units. The NSDI dzongkhag record declares EPSG:5266; its gewog record declares EPSG:4326. Transform coordinates deliberately and verify the delivered files' dates, attributes and topology.

Estimate exposure and interpret risk

Define hazard as the potentially damaging process, exposure as the people or assets within its reach, and vulnerability as their susceptibility to harm and capacity to respond. A scenario-based exposure calculation is:

Here, N_c is population count in grid cell c, and f_c,s is the fraction allocated to the affected area under scenario s. Area-weighted allocation assumes uniform population within a cell. Building-based allocation requires assumptions about residential use and occupancy. State whether the estimate concerns residents, workers, daytime users or seasonal occupants.

WorldPop's Bhutan 2024 record provides constrained population estimates at three arc seconds, approximately 100 m at the equator. Its R2025A release is labelled alpha and subject to revision. Record the exact release, compare totals with the relevant official population and investigate settlements omitted by the allocation mask. WorldPop product record

Rinzin and colleagues (2026) combine hydrodynamic glacial-lake flood scenarios with buildings, transport, farmland and hydropower infrastructure. Population exposure is estimated by allocating each local unit's 2017 census population equally across its mapped buildings. Their relative prioritisation index depends on building completeness, occupancy assumptions and vulnerability indicators. The reported scenarios support comparisons under the study's assumptions; annual expected losses require event probabilities and suitable damage functions. Rinzin et al., 2026

Connect maps with environmental and social research

For hydropower, trace upstream and downstream pathways alongside the intake, dam, powerhouse, access roads and spoil areas. Examine flood exposure, altered flows, sediment movement and resource access. Verify water users and livelihood dependencies through field research, and assess interactions with other infrastructure and operating scenarios.

For roads, compare alternatives using terrain, geology, settlements, habitat and a validated transport network. Estimate travel time and reliability with walking routes, bridges, seasonal closures and detours. Facility operating capacity, transport costs, land acquisition and roadside enterprise effects complete the accessibility assessment.

For conservation and pastoral livelihoods, combine Tsamdro records with participatory mapping of grazing, water, seasonal movement and access arrangements. Establish who uses resources, when and under which relationships. Record-holder and user populations may differ. Protect sensitive locations and community knowledge through agreed disclosure arrangements.

For urban development and tourism, examine housing, service capacity, labour demand and supplier networks. Include renters, temporary workers and informal enterprises in the research frame. Assess benefits and losses for each affected group, with consistent price dates, restoration periods and measures of net income or production.

Preserve comparability and uncertainty

The national land-cover report describes 13 major classes and 87% overall accuracy. Validate the classes important to a decision, including agriculture and small habitat patches, against suitable reference evidence. A land-cover category describes a mapped surface; livelihood use and ecological significance require additional observations. NLCS technical report

Time-series analysis needs compatible classifications and observation periods. ESA explicitly identifies algorithm changes between WorldCover 2020 and 2021. CHIRPS v3 daily products partition pentadal rainfall totals; specify the chosen daily product when examining extremes. ERA5-Land's native 9 km grid supplies regional context, while station and field evidence resolve local conditions. CHIRPS documentation, ERA5-Land metadata

Use categorical resampling for land-cover labels, conserve totals when redistributing population counts and retain survey design information for weighted estimates. Keep collection dates, model years, publication dates and archive dates separate. Products sharing census inputs or satellite imagery also share potential sources of error.

When estimating project effects, establish baseline and follow-up outcomes, comparison populations and relevant spillovers. Difference-in-differences requires a defensible parallel-trends assumption. Matching addresses observed differences; interviews and process evidence help assess mechanisms and remaining confounding. The sampling and quantitative analysis guide and spatial baseline guide develop these design choices.

Curate a reproducible research package

Retain a source register, geographic crosswalk, selected tables or spatial clips, processing recipe and aggregate outputs. Record custodian, edition, observation period, geographic support, units, licence and transformations. Existing archived files should be matched to their original metadata before reuse; upload dates and filenames provide only discovery clues.

Current ICIMOD regional records declare CC BY 4.0, while older packages can carry different agreements. NCHM's data-sharing guidelines require consent for third-party redistribution of supplied records. Apply each edition's terms to both public outputs and authenticated archive access.

Source links lead to providers. The archive table identifies catalogue captures, reports and analytical datasets separately, with access to reviewed editions governed by reader permissions and recorded reuse rights.