WATER CRISIS SCORE
· 51.8% ·
10.4
/ 20
↓
-0.46 since yesterday
ELEVATED
SHARE TODAY'S ASSESSMENT
MODEL BREAKDOWN
ChatGPT
11.0
- Ogallala aquifer depletion threatens U.S. agricultural productivity and rural water access.
- India-Bangladesh tensions over Ganges-Brahmaputra water sharing remain unresolved.
- Himalayan glacial retreat reduces water supply for millions in South Asia.
- UN Watercourses Convention lacks enforcement mechanisms for non-signatory states.
- Urban water infrastructure in Lagos struggles with rapid population growth and inequality.
Claude
14.0
- Ogallala Aquifer depletion rate accelerating, 30% volume loss since 1950
- Ganges-Brahmaputra basin facing 50% flow reduction from Himalayan glacial retreat
- India-Pakistan water treaty under extreme stress from climate change impacts
- Arabian Peninsula groundwater reserves approaching total extraction within 20 years
- Urban water infrastructure failure imminent in multiple Middle Eastern megacities
Gemini
N/A
Llama
14.0
- Ogallala Aquifer depletion exceeds 30% capacity
- Arabian Aquifer over-extraction threatens regional stability
- Ganges-Brahmaputra basin faces critical groundwater decline
- Himalayan glacial retreat accelerates downstream water scarcity
- Transboundary water treaties under increasing diplomatic strain
DeepSeek
13.0
- Ogallala Aquifer depletion accelerating beyond recharge rates, threatening US breadbasket irrigation
- Indus Water Treaty violations escalating as India/Pakistan divert more water for hydropower
- Glacial retreat reducing dry-season flows for Kabul, Islamabad, Kathmandu by 25-40%
- UN Watercourses Convention lacks enforcement mechanisms as Nile dam disputes intensify
- Mexico City sinking 50cm/year from aquifer over-extraction while pipes lose 40% water
Mistral
N/A
Qwen
11.0
- Ogallala Aquifer depletion accelerated to 3.5 km³/year average 2020–2023, exceeding 2010–2019 rate by 27%
- Indus Waters Treaty arbitration mechanisms suspended since 2022 after India revoked Jammu and Kashmir’s special status
- Himalayan glacier mass loss reached 267 Gt/year 2020–2023, reducing dry-season Ganges flow by 12% versus 1990–2010 baseline
- UN Watercourses Convention ratified by only 44 states, with zero enforcement actions taken since 2014
- Jakarta’s groundwater extraction caused 40% of city area to subside >2.5 m since 2010, disabling 37% of piped water infrastructure
Gemma
11.0
- Ogallala Aquifer depletion exceeds recharge in key agricultural zones, accelerating.
- Mekong River Commission lacks enforcement power amid Chinese dam operations.
- Himalayan glacier melt increasingly erratic, impacting Indus basin predictability.
- UN Watercourses Convention ratification remains limited, hindering broad application.
- Rapid urbanization in South Asia strains already-limited groundwater resources.
Nvidia
0.0
Arcee
N/A
Grok
N/A
Perplexity
12.0
- GRACE satellite data show sustained depletion in Ogallala, North China Plain, Northwest India aquifers.
- Arabian Peninsula fossil aquifers heavily overdrawn for irrigation, with minimal natural recharge and substitution.
- Indus and Ganges-Brahmaputra basins already experience high groundwater stress alongside rapid population growth.
- Himalayan and Andean glacier retreat reduces dry-season flows for major Asian and South American rivers.
- Nile, Euphrates-Tigris, Indus basins show recurring diplomatic crises over new dams and allocations.
Cohere
8.0
- Ogallala Aquifer depletion accelerating, losing 1 ft/year in high-use zones
- Indus Basin Treaty under strain from India-Pakistan military posturing
- Himalayan glacial retreat reducing dry-season Ganges flow by 20%
- UN Watercourses Convention non-ratification by key riparian states
- Sub-Saharan urban water access gap widening, 40% unserved in megacities
Jamba
9.0
- Ogallala Aquifer depletion rates exceed 1 foot annually in key regions.
- Ganges-Brahmaputra basin groundwater overuse reduces flow for 400 million people.
- Arabian Aquifer sustains 60% loss, threatening food security in Gulf states.
- Indus Water Treaty frays amid India-Pakistan military escalations over river flow.
- Glacial retreat in Andes and Himalayas cuts dry-season flow for 2 billion downstream.
Moonshot
11.0
- Ogallala aquifer depletion accelerated 6% 2022-2024, Kansas irrigated acreage down 30%
- Saudi Arabia ceased own wheat farming 2016, now draining Arizona aquifers via land purchases
- Indus Waters Treaty arbitration stalled since 2022, India withholding hydrological data from Pakistan
- Tibetan Plateau glaciers lost 15% mass 2000-2023, reducing dry-season flow to 1.8B people
- Lagos water utility serves 10% of 22M residents, remainder depends on contaminated boreholes
AI DISAGREEMENT ANALYSIS
3.72
STD DEV
DIVIDED
BASED ON 11 MODEL RESPONSES
The models' score divergence stems from differential weighting of key hydrological stress indicators, with higher-scoring models like Claude and Llama emphasizing accelerated aquifer depletion rates and transboundary water treaty tensions, particularly in the Ogallala, Arabian, and Indus basins. Lower-scoring models such as Cohere and Jamba applied more conservative thresholds for water security risk, focusing on incremental changes in glacial retreat and urban water access rather than systemic infrastructure vulnerabilities. Methodological variations emerged in how models integrated climate-driven flow reduction, diplomatic tension metrics, and regional water stress aggregation, with Claude and Llama demonstrating more holistic risk modeling by synthesizing glacial retreat, groundwater extraction rates, and geopolitical friction indicators across multiple water-stressed regions.
AI-GENERATED ANALYSIS — NOT AN OFFICIAL ASSESSMENT
ABOUT THIS SYSTEM
The Water Crisis Score (WCS) is an experimental daily assessment
aggregated from leading AI models.
Each model is independently asked to rate the current global water crisis risk on a scale of 0 to 20
and provide up to 5 specific reasons for its assessment.
The final score is the average of all valid responses.
This is not an official or governmental assessment.
It reflects the synthesised opinion of large language models based on their training and world knowledge.
Visit the [BRIEFING] page for more information.