Roxr Climate Intelligence Index · RCII v2.0

Drought doesn't
announce itself.
We do.

Monthly satellite-driven drought risk intelligence for Kenya's arid and semi-arid lands — 26 years of data, three independent signals, and a 90-day forward view your team can act on before conditions deteriorate.

Live assessment · March 2026
Kenya ASAL
2
At elevated risk
1.8M
Population assessed
10
Counties monitored
90-day
Forecast horizon
Signal to intelligence

Three signals.
One clear picture.

Each month, three independent satellite data streams are ingested, validated against historical baselines, and fused into a single county-level risk score. If one signal misfires, the others catch it. A genuine drought registers across all three simultaneously.

01 · Vegetation
MODIS NDVI1 km monthly2000–present
Vegetation Health Anomaly

MODIS Terra measures the ratio of near-infrared to red light reflected by the land surface — a precise measure of how much green biomass is actively photosynthesising. We compute the monthly anomaly against the 2000–2015 climatological baseline. Browning below normal is one of drought's earliest fingerprints, detectable weeks before ground reports confirm the situation.

02 · Thermal
MODIS LST1 km 8-day2000–present
Land Surface Temperature

As soil moisture depletes, evaporative cooling fails. Land surfaces run 8–12°C above their seasonal baseline during severe drought. During Kenya's 2011 crisis, northern counties recorded LST anomalies exceeding +10°C before food insecurity was officially declared. MODIS 8-day composites are aggregated to monthly and quality-masked to remove cloud-contaminated pixels.

03 · Precipitation
CHIRPS SPI-3~5 km pentad1981–present
Standardised Precipitation Index

The 3-month SPI measures accumulated rainfall deficit against a 1981–2010 baseline using Gamma distribution fitting — the WMO standard methodology. SPI-3 captures whether Kenya's bimodal rainfall seasons are delivering sufficient moisture to sustain pastoral and agricultural systems, with the 3-month window matching drought onset timing in ASAL counties.

How the signals become a score. Each signal anomaly is normalised within each county's own seasonal history — Turkana in dry season is assessed against Turkana's dry-season record, not against wetter counties. The three normalised stress values are fused using seasonally-adjusted weights and multiplied against a county-level Vulnerability-Exposure Index (population, livestock, water access, cropland dependency, adaptive capacity). The result is the RCII and the RRIS — both on a 0–1 scale.

RCII = f(veg_stress, thermal_stress, precip_stress) × seasonal_weights · RRIS = RCII × VEI × forecast_multiplier
Monthly deliverables

Intelligence products
ready to distribute.

Every monthly run produces three outputs from the same underlying risk data. Your PDF briefing, your dashboard, and your data feed always tell the same story — because they come from a single validated pipeline.

📄
5-Page PDF Intelligence Brief

Executive summary with KPI tiles and AMBER-tier alert cards. Full county risk matrix. Signal decomposition charts showing what is driving each county's score this month. 25-year historical RRIS trends with documented drought event markers. Full methodology note. Designed to forward directly to boards, government counterparts, and field partners.

PDF5 pagesMonthlyClient-branded
📊
Structured Data Feeds

Machine-readable JSON and CSV with RRIS scores, risk tiers, 30/60/90-day forecast classes, confidence levels, affected population figures, and decision trigger text for all 10 counties. Validated against 14 schema checks before release. Consumable by risk platforms, insurance pricing engines, and early warning dashboards via direct API integration.

JSONCSVAPI-ready14-point validation
🗺
Interactive HTML Dashboard

A browser-based choropleth map of all 10 ASAL counties colour-coded by risk tier. Click any county for its full risk profile: RRIS, RCII hazard, VEI vulnerability, forecast table, affected population, and decision trigger. Embeddable in client portals, shareable as a 96 KB standalone file, readable on any device without installation.

HTMLInteractive mapEmbeddable96 KB
26
Years of training data
2000–2026 · extended annually
3
Independent satellite signals
NDVI · LST · SPI-3 · cross-validated
90d
Maximum forecast horizon
30d · 60d · 90d classifiers
Risk tiers and decisions

Risk intelligence that
moves decisions forward.

The RCII is built for organisations that must act before a drought crisis peaks. Each risk tier carries a specific decision mandate — not just a colour, but a clear instruction.

RED
0.45 – 0.65
Activate response. Livelihood support and emergency feeding required. Escalate to inter-agency coordination.
AMBER
0.25 – 0.45
Prepare interventions. Pre-position resources, activate index-insurance trigger protocols, engage county partners before conditions deteriorate.
YELLOW
0.10 – 0.25
Enhanced monitoring. Elevate monitoring frequency, refresh logistics pre-positioning, brief rapid-response teams.
GREEN
0.00 – 0.10
Routine monitoring. Conditions within normal parameters. No active resource mobilisation required.
1,804,513
People in the March 2026 assessed population across 10 Kenya ASAL counties
Live assessment · 2 counties at AMBER
0.547
30-day model balanced accuracy — 2.7× above the random baseline for a 5-class forecast problem
TimeSeriesSplit CV · 3,060 training samples
48h
Maximum delivery lag from MODIS data availability to client brief in your inbox
Monthly cycle · fully automated pipeline
2011
East Africa drought identified from satellite signals alone — Mandera reached RRIS 0.828, consistent with documented NDMA emergency declarations
Historical validation · no ground data required
Who we work with

Built for organisations
that act early.

The value of drought intelligence is front-loaded. The organisations that benefit most are those that make decisions weeks before a crisis becomes visible on the ground.

🌾
Agricultural Insurers

Activate index-insurance triggers at the right moment. The RCII gives you a 90-day view on which counties are approaching trigger conditions — ahead of fieldwork season and claims peaks.

🏦
Development Finance

Flag climate-exposed loan portfolios before borrowers default. County-level RRIS scores map directly to agricultural and pastoral livelihood exposure across ASAL portfolio geographies.

🚁
Humanitarian Agencies

Move from reactive to anticipatory. Monthly briefings let you pre-position logistics and brief rapid-response teams while there is still time to act cost-effectively — before emergency declarations.

🏛
Government & NDMA

A consistent, satellite-based early warning signal that complements ground-based monitoring. Monthly county briefings ready to share with county drought management committees.

Get in touch

See the March 2026
intelligence brief.

We'll send you the complete demonstration package — the same PDF, data feed, and interactive dashboard your team would receive every month.

Monthly SubscriptionPDF + feeds + dashboard · delivered within 48h
Custom IntelligencePortfolio-specific briefs · client-branded outputs
Data API AccessDirect JSON integration into your risk systems
intelligence@roxr.earth

Kenya ASAL · Eastern Africa · RCII v2.0 · roxr.earth