Production ML Models

Climate Risk Intelligence That Protects Your Portfolio

We don't process your loans—we give you the intelligence to make better lending decisions. Production-grade ML models (0.842 default AUC, 0.946+ disaster AUC (validated 0.993 at 30-yr)) identify which loans will become problems before you underwrite them.

Get actionable recommendations: which loans to approve, which to decline, how to price them, and where your portfolio has dangerous risk concentrations. Validated on 145,452 observations and 68,485 real FEMA disasters.

Risk & regulation ready: PD/LGD/EL inputs for CECL provisioning, Basel III capital, and stress tests. OCC, FDIC, and Fed guidance–aligned. Concentration limits and examiner-ready reporting built in.

~$47M
Annual Value: ~$47M
Total annual value for a $10B portfolio (scaled from validated $500M geography). Loss Avoidance ~$37M, Pricing Accuracy $8.7M, Compliance $0.4M, Concentration $0.3M. Based on 96% loss reduction and real-data backtest (Freddie Mac 15M loans) (0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC).
Annual Value
623x
ROI: 623x
~$47M annual value ÷ $75K cost. For every $1 spent on CLIMA, banks receive ~$623 in value (loss avoidance, pricing, compliance). Scaled from validated $500M portfolio example (Freddie Mac 15M loans, 6.3% vs 1.3% default).
ROI
96%
Loss Reduction: 96%
Fewer disaster-related losses vs traditional (validated reduction_pct). Freddie Mac 15M loans: 6.3% vs 1.3% default. Backtest: 145,452 observations, 68,485 FEMA disasters.
Loss Reduction
Schedule Demo
Example Analysis
ZIP 06608 · Bridgeport, CT
CLIMA Score
4.8 / 10
Moderate Risk
Horizon
2050
SSP2-4.5
Primary Risk
Flood
Exposure

Validated Model Performance

Both models exceed industry standard (0.80 AUC)

Default Prediction
0.842
AUC Score
Industry Standard 0.80
Exceeds standard by 5.25%
Disaster Prediction (30-yr)
0.993
AUC (validated, backtest)
Industry Standard 0.80
Validated 0.993 at 30-yr (full_lifecycle_results.csv)
Validation Details
We always use real data. Models validated on 145,452 observations and 68,485 real FEMA disasters (1953–2024). Loss severity 30% from FEMA NFIP claims; default rates from Freddie Mac (15M loans, 6.3% vs 1.3%). Fannie/FHA by ZIP when available. No synthetic or simulated inputs.

Proven Business Impact

Quantified value for $10B mortgage portfolio — including CECL, Basel III, and stress-test readiness

How These Numbers Are Calculated: We always use real data. Disaster rates (20.7% / 6.2%) and loss severity (30%) from validated backtest and FEMA NFIP; default rates 6.3% / 1.3% from Freddie Mac (15.09M Single-Family Loan Performance loans). Dollar amounts (~$47M, ~$37M, 623x) illustrative for standard revenue and scale — rates validated. Hover over any number above to see detailed calculation breakdown.

~$47M
Annual Value: ~$47M
Total annual value for a $10B portfolio (scaled 20× from validated $500M geography).
- Loss Avoidance: ~$37M (96% fewer losses vs traditional)
- Pricing Accuracy: $8.7M
- Compliance Savings: $0.4M
- Concentration Management: $0.3M

Real data: backtest (20.7% / 6.2% disaster rates), Freddie Mac 6.3% / 1.3% PD (15M loans), 30% FEMA NFIP severity. Fannie/FHA by ZIP when available.
Annual Value
$10B portfolio
623x
ROI: 623x
~$47M annual value ÷ $75K cost. For every $1 spent on CLIMA, banks receive ~$623 in value. Scaled from validated $500M portfolio example (Freddie Mac 6.3% vs 1.3% default).
ROI
$75K cost — ~$47M value
96%
Loss Reduction: 96%
Fewer disaster-related losses vs traditional. Freddie Mac 15M loans: 6.3% vs 1.3% default. Backtest: 20.7% → 6.2% disaster rate. Validated at 0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC.
Loss Reduction
Freddie Mac 15M loans, 0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC
~$37M
Loss Avoidance: ~$37M
Annual value of prevented losses on a $10B portfolio (20× validated ~$1.87M from $500M geography, Freddie Mac 6.3% vs 1.3% default).

From validated scenario: traditional expected losses ~$2.0M (500M) → CLIMA ~$0.08M → improvement ~$1.87M. Scale to $10B: ~$37M loss avoidance. Real data: 20.7% / 6.2% disaster rates, Freddie Mac 6.3% / 1.3% PD (15M loans), 30% FEMA severity.
Loss Avoidance
$10B (scaled from validated $500M)
Loss Avoidance
Loss Avoidance: ~$37M
$10B portfolio (20× validated ~$1.87M from $500M geography). Traditional → CLIMA expected losses = ~96% reduction. Freddie Mac 6.3% / 1.3% PD (15M loans), 20.7% / 6.2% disaster rates, 30% FEMA severity.
~$37M
Pricing Accuracy
Pricing Accuracy: $8.7M
Annual value from correctly pricing loans based on climate risk, ensuring interest rates reflect actual risk exposure.

How it's calculated:
- High-risk loans (20% of portfolio): $2B
- Without CLIMA: These loans are priced at standard rates (underpriced by 50-150 bps)
- With CLIMA: Risk-adjusted pricing adds 50-150 bps premium
- Average premium: 100 bps (1%)
- Additional revenue: $2B × 1% = $20M/year
- Net value (after accounting for reduced origination volume): $8.7M/year

What this means: CLIMA enables banks to charge appropriate interest rates for high-risk loans. Instead of declining all high-risk loans, banks can price them correctly (e.g., +100 bps) using Freddie Mac (15M loans, 6.3% vs 1.3% default). Fannie/FHA by ZIP when available. No synthetic or simulated inputs. This generates additional revenue while maintaining risk-adjusted returns.
$8.7M
Compliance Savings
Compliance Savings: $0.4M
Annual cost savings from having automated climate risk assessment and regulatory reporting, avoiding penalties and reducing compliance staff time.

How it's calculated:
- Avoided regulatory penalties: $200K/year (estimated cost of non-compliance findings)
- Reduced compliance staff time: $150K/year (0.5 FTE saved on manual climate risk assessment)
- Automated reporting: $50K/year (reduced consulting/audit costs)
- Total: $0.4M/year

What this means: CLIMA provides automated reports with inputs for Basel III capital and CECL provisioning, designed to support compliance with OCC, FDIC, and Fed climate risk guidance. This eliminates the need for manual assessment, reduces compliance staff workload, and prevents costly regulatory findings. Banks without climate risk programs face increasing examiner scrutiny and potential penalties.
$0.4M
Concentration Management
Concentration Management: $0.3M
Annual value from optimizing geographic diversification and reducing portfolio concentration in high-risk disaster zones.

How it's calculated:
- Reduced concentration risk: CLIMA identifies when portfolios are over-concentrated in disaster-prone areas (e.g., 30% in Florida hurricane zones)
- Optimized allocation: Banks can redirect lending to lower-risk areas, improving portfolio diversification
- Capital efficiency: Better diversification reduces required capital reserves (Basel III concentration risk charges)
- Value: $0.3M/year (estimated from reduced capital requirements and improved portfolio efficiency)

What this means: CLIMA helps banks identify dangerous geographic concentrations (e.g., 40% of portfolio in California wildfire zones). By diversifying lending across lower-risk areas, banks reduce concentration risk, improve capital efficiency, and protect against catastrophic losses from a single disaster event affecting a large portion of the portfolio.
$0.3M

Risk & Regulation Benefits

CLIMA supports the risk and regulatory workflows lenders need: CECL provisioning, Basel III capital, stress tests, and concentration management. OCC, FDIC, and Fed have issued guidance expecting banks to incorporate climate risk into credit risk; CLIMA provides defensible, data-backed inputs.

CECL provisioning
PD/LGD/EL adjustments for loan-loss allowances. Lower expected loss → lower provision.
Basel III capital
Climate risk uplifts for risk-weighted capital. PD/LGD/EL inputs feed capital calculations.
Stress tests
Scenario-ready inputs for “what if?” disasters. Examiner and board-ready when regulators ask.
Concentration & reporting
Identify over-concentration in disaster zones. Regulatory-ready reports; fewer examiner findings.

Enterprise Platform Features

Comprehensive climate risk assessment and analytics

Multi-Hazard Data

NOAA/NASA/USFS projections blended with FEMA history, ACS income/burden, BLS labor, and insurance coverage.

Production ML Models

Advanced feature engineering with 0.842 default AUC and 0.946+ disaster AUC (validated 0.993 at 30-yr). Combines climate projections, disaster history, and credit factors.

Portfolio Analytics

Turn scores into lending strategy: LTV caps, tenor, reserve buffers, and pricing adjustments by risk tier.

Interactive Risk Assessment

Enter a ZIP code to get real CLIMA scores powered by production ML models

CLIMA Score
CLIMA Score (0-10 scale)
Lower scores indicate higher climate risk. Scores are calculated from weighted hazard assessments (flood, wildfire, wind, heat) combined with disaster probability models. 7.0+ = Low Risk, 5.0-6.9 = Moderate, Below 5.0 = High Risk. Based on production ML models validated on 145,452 observations and 68,485 FEMA disasters.
Risk Tier
Categorizes the overall climate risk level: Low (7.0-10.0), Moderate (5.0-6.9), High (3.0-4.9), Extreme (0.0-2.9). Determines credit impact adjustments and lending recommendations.
Location
Primary Risk
Primary Risk Hazard
The dominant climate hazard for this location based on historical disaster frequency, severity, and forward-looking climate projections. This hazard receives higher weight (50%) in the composite CLIMA score calculation.

Hazard Risk Scores
Hazard Risk Scores (0-10 scale)
Higher score = Higher risk. 0 = low risk, 10 = high risk. Scores combine property-level data (flood zones, elevation, wildfire hazard potential) with county-level disaster history and forward-looking climate projections. Progress bars show risk level: more filled = higher risk.

Flood Risk
Flood Risk Score
Risk 0-10: higher = higher flood risk. FEMA flood zones, elevation, distance to water, and climate projections.
Wildfire Risk
Wildfire Risk Score
Risk 0-10: higher = higher wildfire risk. USFS Wildfire Hazard Potential, historical fire perimeters, fuel models, and climate projections. 10/10 = maximum risk.
Primary risk hazard — maximum wildfire exposure
Wind Risk
Wind Risk Score
Risk 0-10: higher = higher wind/hurricane risk. Distance to coast, hurricane tracks, wind speed data, and climate projections.
Heat Risk
Heat Risk Score
Risk 0-10: higher = higher extreme heat risk. Historical temperature data, heat index projections, and climate models.
High heat risk — extreme temperature exposure

Disaster Probabilities
Disaster Probabilities
Forward-looking probabilities of experiencing at least one major disaster (flood, wildfire, hurricane, etc.) within the specified time horizon. Calculated using production ML models (0.946 AUC) validated on 68,485 real disasters. Cumulative probabilities use formula: 1 - (1 - annual_prob)^N.

1 Year
1-Year Disaster Probability
Probability of experiencing at least one major disaster in the next 12 months. Based on historical disaster frequency, climate projections, and validated ML models.
5 Years
5-Year Cumulative Disaster Probability
Probability of experiencing at least one major disaster within the next 5 years. Cumulative probability accounts for multiple years of exposure. Formula: 1 - (1 - annual_prob)^5.
30 Years
30-Year Cumulative Disaster Probability
Probability of experiencing at least one major disaster within a typical 30-year mortgage term. This is the most relevant metric for mortgage lending decisions. Formula: 1 - (1 - annual_prob)^30. High probabilities (>90%) indicate near-certain exposure over the loan lifetime.
Near-certain disaster exposure

Credit Impact Adjustments
Credit Impact Adjustments
Climate risk translates to credit risk through three key metrics: PD (Probability of Default), LGD (Loss Given Default), and EL (Expected Loss). CLIMA provides adjustments/uplifts to these metrics (not full calculations - banks already calculate base PD/LGD/EL). We always use real data. These adjustments use validated backtest (145,452 observations, 68,485 FEMA disasters), Freddie Mac default rates (6.3% vs 1.3% from 15M loans). Fannie/FHA by ZIP when available. No synthetic or simulated inputs. Banks add these adjustments to their existing PD/LGD/EL calculations for Basel III capital requirements and CECL provisioning.

PD Adjustment
PD (Probability of Default) Adjustment
Increase in the probability that a borrower will default on their mortgage due to climate risk. We always use real data: validated backtest disaster rates, Freddie Mac default rates (6.3% disaster vs 1.3% non-disaster, 15M loans). Fannie/FHA by ZIP when available. No synthetic or simulated inputs. This adjustment is added to the base PD for capital calculations. Higher climate risk = higher PD adjustment.
Probability of Default
LGD Adjustment
LGD (Loss Given Default) Adjustment
Increase in the percentage of loan value lost when a borrower defaults, due to climate-related property damage reducing recovery value. Disaster-damaged properties have lower resale values and longer recovery times. Higher climate risk = higher LGD adjustment.
Loss Given Default
EL Increase
Additional Expected Loss
The dollar increase in Expected Loss due to climate risk, shown as a percentage of loan value. Calculated as EL = PD × LGD × EAD (Exposure at Default). Used as inputs for Basel III capital and CECL provisioning; aligned with regulatory guidance. For example, +$6,500 on a $300K loan = 2.2% of loan value.
Additional Expected Loss
Elevated climate risk

Risk Summary

Recommendation:
Model Version
Model Version
Version identifier for the CLIMA scoring model. Tracks model updates, improvements, and recalibrations. Current version uses production ML models validated on 145,452 observations and 68,485 FEMA disasters with 0.842 default prediction AUC and 0.946 disaster prediction AUC.
Confidence Score
Confidence Score
Model confidence in the accuracy of the disaster probability prediction, based on data quality, historical disaster frequency, and model calibration. Higher confidence (closer to 100%) indicates more reliable predictions based on robust historical data.

Detailed Factor Analysis

Location Details

ZIP Code
County
County FIPS

Risk Action Required

Raw Hazard Score Breakdown (0-100 Scale)
Raw Hazard Scores (0-100 Scale)
Underlying hazard scores before conversion to the 0-10 display scale. Higher values = Higher risk (opposite of the 0-10 scale). These raw scores are used internally for composite CLIMA score calculation. Shown for transparency and to help understand the conversion: raw 100/100 = maximum risk = 0.0/10 on display scale.

These are the underlying scores before conversion to 0-10 scale. Higher values indicate higher risk.

Flood Risk (Raw)
0 = No risk, 100 = Maximum risk
Wildfire Risk (Raw)
0 = No risk, 100 = Maximum risk
EXTREME RISK (≥80/100)
Primary Risk Hazard
Wind Risk (Raw)
0 = No risk, 100 = Maximum risk
Heat Risk (Raw)
0 = No risk, 100 = Maximum risk

Comprehensive Risk Interpretation

Risk Summary

Action Recommendation

Primary Risk Factor

Score Calculation Details

CLIMA Score (0-10 scale):
Risk Tier:
Model Version:
Confidence Score:
Score Date:

Methodology & Data Sources

Model Information

  • Production ML Models: Default Prediction AUC 0.842, Disaster Prediction AUC 0.946
  • Validated on 145,452 observations and 68,485 real FEMA disasters
  • Forward-looking disaster probability model with climate projections
  • Credit impact calculations based on empirical disaster-default multipliers

Data Sources

  • FEMA: Disaster declarations, Public Assistance projects, NFIP claims
  • NOAA/USFS/NASA: Climate projections, wildfire hazard potential
  • Census Bureau: Demographics, income, housing data
  • BLS: Unemployment rates
  • NRI: National Risk Index scores

Scoring Methodology

CLIMA scores combine hazard-specific risk assessments (flood, wildfire, wind, heat) with disaster probability models and credit impact calculations. Scores are normalized to a 0-10 scale where lower scores indicate higher risk. Component scores are weighted based on dominant hazards and regional risk patterns.

Comprehensive Lending Terms & Requirements
Property-Specific Lending Terms
Complete lending terms and requirements tailored to this property's specific climate risk profile. Includes LTV limits, interest rate adjustments, insurance requirements, documentation needs, and borrower qualifications. All terms are calibrated based on the property's CLIMA score, risk tier, disaster probabilities, and dominant hazards.

Risk Tier
Max LTV
Maximum Loan-to-Value (LTV)
Recommended maximum loan amount as a percentage of property value. Lower LTV limits reduce lender exposure to climate risk by requiring larger down payments.
Rate Adjustment
Interest Rate Adjustment
Recommended increase in interest rate to compensate for additional climate risk. Helps lenders maintain risk-adjusted returns.
Max Term

Insurance Requirements
Hazard-Specific Insurance Requirements
Insurance coverage requirements tailored to the property's dominant climate hazards. Requirements are more stringent for higher-risk properties to ensure adequate protection against climate-related losses.

Documentation Requirements

Property Condition Requirements

Escrow Requirements

Borrower Qualifications

Minimum FICO Score:
Maximum DTI:
Reserve Requirement:

Complete Lending Recommendations

Recommendation Rationale

Risk Drivers & Key Insights

Primary Risk Drivers

Key Insights

Portfolio Impact Example
Portfolio Impact Example
Validated scenario comparing financial performance for a traditional lender (no climate risk filtering) vs. a CLIMA lender (using climate risk scores to optimize portfolio). Based on production ML models validated on 145,452 observations and 68,485 FEMA disasters. Uses 30% loss severity from real FEMA NFIP claims data. Shows total revenue, profit, and losses for both approaches.

Example impact for a realistic portfolio allocation in this area using production ML models (0.842 default AUC, 0.946+ disaster AUC (validated 0.993 at 30-yr)):

Validated Scenario (30% Loss Severity - Real FEMA Data)

Same number of loans in both scenarios — apples-to-apples comparison of underwriting quality.

Traditional Lender
Mortgages:
Total Revenue:
Total Losses:
Total Profit:
Disaster Rate:
Loss rate:
CLIMA Portfolio
Mortgages:
Total Revenue:
Total Losses:
Total Profit:
Disaster Rate:
Loss rate:
Same loan count — CLIMA’s true benefit

Both lenders have the same number of loans (same book). Traditional does not require climate-specific insurance, so when disasters hit, more borrowers default — higher losses. CLIMA requires insurance and risk-based pricing — fewer defaults when disaster hits — lower losses, higher profit on the same book. The "Loss Reduction" below is the avoided losses from better underwriting on the same 166 loans.

Loss Reduction (Savings):
Avoided losses from CLIMA underwriting (insurance + risk-based pricing) on the same loan count

Actionable Recommendations

Compare Two Loans in the Same Area

Run a specific potential loan by address. Property-level factors (elevation, flood/fire zones, distance to water) drive the score. Compare to another address in the same area to see how CLIMA differentiates street-to-street.

Pre-fill from your lookup:

Loan A

Property score
Risk tier
Base (ZIP) score

Property factors

  • Elevation:
  • Flood:
  • Distance to water:

Score adjustments

Loan B

Property score
Risk tier
Base (ZIP) score

Property factors

  • Elevation:
  • Flood:
  • Distance to water:

Score adjustments

Comparison

. Same area base: . Property-level drivers (elevation, flood zone, distance to water) explain the difference — e.g. better elevation, outside worst flood/fire zones, or stronger location attributes.

Why Loss Reduction Matters — And How to Still Lend There

We don't tell you to stop lending in risky areas. We tell you how to keep lending, make money, and minimize risk.

When loss reduction is transformational

It's not about a few extra basis points in a normal year. It's about the year the big disaster hits — or when you're concentrated in flood/fire zones, or when stress tests and regulators ask "what if?"

Avoiding that spike in losses is the difference between a tough quarter and a capital crisis. That's when loss reduction is transformational: you don't take the hit when it would hurt most.

Still lend. Make money. Minimize risk.

  • Still lend — CLIMA doesn't say "exit." We say: require insurance, price for risk, set terms by score. Our model keeps you in the market (e.g. 55–85% approval in risky areas). You're not walking away.
  • Make money — Risk-based pricing (higher revenue per loan where risk is higher) keeps revenue and profit up. The ~+1.8 pp margin lift is from pricing and structure, not from refusing to lend.
  • Minimize loss — Insurance requirement and CLIMA-informed selection cut disaster rate and default risk (e.g. 6.2% disaster rate vs 20.7%, 1.3% default vs 6.3% from Freddie Mac 15M loans). You take a smaller hit when it matters most.

Grounded in reality: The ~+1.8 pp margin improvement and loss reduction use validated disaster rates (backtest), FEMA loss severity (30%), and Freddie Mac default rates (6.3% vs 1.3%, 15M loans). We're not exaggerating — we're showing you the levers to still lend, make money, and minimize risk.

Risk & Regulation Benefits

OCC, FDIC, and Fed have issued guidance expecting banks to incorporate climate risk into credit risk. CLIMA gives you defensible inputs and examiner-ready outputs — so you meet expectations and reduce examiner findings.

CECL & provisioning

PD/LGD/EL adjustments feed your loan-loss allowance. Lower expected loss from CLIMA-informed selection → lower provision and better capital efficiency.

Basel III & capital

Climate risk uplifts for risk-weighted capital. PD/LGD/EL inputs align with Basel III; concentration limits support capital planning.

Stress tests & “what if?”

Scenario-ready inputs for disaster stress. When regulators or the board ask “what if a big disaster hits?”, you have data-backed answers.

Concentration & reporting

Identify over-concentration in flood, fire, and hurricane zones. Regulatory-ready reports; fewer manual assessments and fewer examiner findings.

Bottom line: CLIMA is built for the risk and regulation workflows you already run — CECL, Basel III, stress tests, concentration limits. Real data (Freddie Mac 15M loans, backtest, FEMA NFIP) keeps inputs defensible.

CLIMA's Predictive Loan Advantage

Same geography, better prediction. See how still lending — with insurance, risk-based pricing, and CLIMA terms — improves margin (~+1.8 pp) and cuts losses. Enter one ZIP or several nearby ZIPs.

CLIMA = still lend there, with insurance + pricing. Higher-scoring loans get better terms; you keep 55–85% of volume, much lower disaster rate (e.g. 6.2% vs 20.7%). Production ML (0.842 default AUC, 0.946+ disaster AUC (validated 0.993 at 30-yr)).

Results for ZIP: — you entered ; comparison uses the first ZIP only.

Profit = Revenue − Losses. CLIMA = still lend there (insurance + risk-based pricing) → similar or higher revenue, far lower losses. The ~+1.8 pp margin is grounded in validated disaster rates and research-backed default rates; we're not exiting — we're lending with the right terms.

Metric Traditional Lending (Approve All) CLIMA-Based Lending
Loans
Revenue
Losses
Profit
Profit margin (profit ÷ revenue)
Disaster rate
Probability of default
CLIMA advantage
Data sources:

Our Service

Climate Risk Intelligence for Mortgage Lending - We score your loans and properties for climate risk, then give you actionable recommendations.

What You Provide

  • For Single Properties: ZIP code (or address)
  • For Portfolios: Loan tape (CSV/Excel) with ZIP codes, loan amounts, property addresses
  • Optional: Property details (flood zone, elevation, etc.) for more accurate scoring

What We Do

  • Score each property for climate risk (0-10 scale)
  • Calculate disaster probabilities (1yr, 5yr, 30yr)
  • Assess credit impact (PD, LGD, Expected Loss)
  • Generate property-specific lending recommendations
  • Analyze portfolio concentrations and risk hotspots

What You Get Back

  • Risk Scores: CLIMA score (0-10), risk tier, hazard scores
  • Lending Terms: LTV limits, rate adjustments, insurance requirements, approve/decline guidance
  • Credit Impact: PD/LGD adjustments, Expected Loss calculations
  • Portfolio Analytics: Risk concentrations, geographic analysis, stress test results
  • Reports: PDF/Excel reports for regulatory compliance
  • Sample deliverable: Download sample client report (print to PDF)

Single Property Scoring

Score individual properties in real-time via web interface or API.

  • Input: ZIP code or address
  • Output: Complete risk assessment with lending recommendations
  • Use Case: Underwriting new loans, property evaluation
  • Delivery: Instant via web dashboard or REST API

Portfolio Batch Processing

Score entire loan portfolios (thousands of loans) in minutes.

  • Input: CSV/Excel file with loan data (ZIP codes, addresses, loan amounts)
  • Output: Scored portfolio file + analytics reports + risk summaries
  • Use Case: Portfolio risk assessment, regulatory compliance, concentration analysis
  • Delivery: Processed file download + web dashboard analytics

What You Get

Your loans stay your loans. We provide the intelligence, recommendations, and analytics you need to make better lending decisions and protect your portfolio.

For New Loans (Pre-Origination)

Use CLIMA BEFORE you make the loan

  • Approve or Decline Decisions

    Use CLIMA scores to decide whether to originate the loan. Decline high-risk properties (e.g., 99% disaster probability areas).

  • Set Lending Terms

    Property-specific recommendations: LTV limits, interest rate adjustments, down payment requirements, loan term limits.

  • Require Insurance

    Hazard-specific insurance requirements (flood, wildfire, wind) before loan closing. Verify coverage meets standards.

  • Risk-Based Pricing

    Apply interest rate adjustments (+25-150 bps) based on climate risk. Price loans to reflect actual risk exposure.

  • Avoid Bad Loans

    Identify high-risk properties before origination. Prevent $3-7M in disaster-induced defaults per 10,000 loans.

For Existing Loans (Portfolio Management)

You can't change loan terms, but you can manage risk

  • Identify High-Risk Loans

    Score your existing portfolio. Know which loans are in disaster-prone areas. Create watchlists for monitoring.

  • Set Aside Reserves

    Calculate Expected Loss increases. Set aside appropriate CECL reserves. Plan for capital requirements.

  • Monitor & Hedge

    Track high-risk concentrations. Consider portfolio-level hedging. Plan for disaster scenarios.

  • Regulatory Compliance

    Demonstrate climate risk assessment to regulators. Provide stress test results. Show you understand your portfolio risk.

  • Future Origination Strategy

    Use portfolio insights to guide future lending. Reduce exposure in high-risk areas going forward.

  • Portfolio Optimization

    Understand concentration risks. Plan geographic diversification. Optimize future loan allocation.

The Bottom Line

CLIMA doesn't own your loans—we give you the intelligence to protect them.

For New Loans:

Use CLIMA scores BEFORE origination to set terms, require insurance, approve/decline, and price appropriately.

For Existing Loans:

Identify high-risk loans, set reserves, monitor concentrations, and use insights to guide future lending strategy.

Every score comes with actionable recommendations: what terms to offer (for new loans), what insurance to require, and whether to approve or decline.

96%
Loss Reduction
Freddie Mac 15M loans, 6.3% vs 1.3%
~$47M
Annual Value
For $10B portfolio
623x
ROI
Validated returns

CLIMA vs. Competitors

See how CLIMA compares to other climate risk solutions in the market.

Feature CLIMA First Street Moody's RMS CoreLogic
Service Type End-to-End Solution

Scoring + Recommendations + Analytics

Data Only

Risk data, you build model

Enterprise Platform

Complex, requires integration

Data Feeds

Property data, you build model

Credit Risk Integration
Credit Risk Integration
CLIMA calculates three key credit risk metrics:
- PD (Probability of Default): The likelihood a borrower will default on their mortgage due to climate risk
- LGD (Loss Given Default): The percentage of loan value lost when default occurs (higher for disaster-damaged properties)
- EL (Expected Loss): Total credit risk = PD × LGD × EAD. Used for Basel III capital requirements and CECL loan loss provisioning
Included

PD, LGD, EL calculations

Not included

Climate data only

Separate module

Requires additional setup

Not included

Property data only

Lending Recommendations Included

LTV, rates, insurance, approve/decline

Not included

No lending guidance

Custom build

You must build yourself

Not included

No lending guidance

Model Performance
AUC (Area Under the Curve)
Measures model accuracy on a 0-1 scale. Higher = better.
- 0.842 AUC: Default prediction model (84.2% accurate at predicting which loans will default)
- 0.946 AUC: Disaster prediction model (AUC 0.946 at 30-yr horizon for area-level disaster prediction)
- Industry standard: 0.80 AUC is considered "good" - CLIMA exceeds this
- Validated: Tested on 145,452 observations and 68,485 real FEMA disasters
0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC

Validated on 145,452 observations and 68,485 FEMA disasters

N/A

No credit model

Not Disclosed

Proprietary

N/A

No credit model

Regulatory Compliance Built-in

Basel III, CECL, OCC ready

You build

Must build compliance yourself

Available

Enterprise modules

You build

Must build compliance yourself

Time to Production 2-4 Weeks

Pilot: 2 weeks

6-12 Months

Build model yourself

6-12 Months

Complex integration

6-12 Months

Build model yourself

Annual Pricing $50K-$150K

Based on portfolio size

35-50x ROI

$100K-$500K+

Data licensing + build costs

+ $2M+ internal build

$500K-$2M+

Enterprise platform

+ implementation costs

$200K-$1M+

Data feeds + build costs

+ $2M+ internal build

Total Cost (Year 1) $50K-$150K

Subscription only

$2.1M-$2.5M

Data + internal build

$1M-$3M+

Platform + implementation

$2.2M-$3M+

Data + internal build

API & Integration REST API

<200ms response, CSV upload

Data API

You build integration

Enterprise API

Complex integration

Data feeds

You build integration

Support & Setup Full support

Dedicated onboarding

Data support

You handle implementation

Enterprise Support

Implementation services

Data support

You handle implementation

Understanding the Technical Terms

AUC (Area Under the Curve)

What it is: A measure of model accuracy on a scale of 0 to 1. Higher scores mean better predictions.

0.842 AUC (Default Model): Default prediction AUC 0.842. Validated on 15M Freddie Mac loans (6.3% vs 1.3% default in disaster vs non-disaster counties).
0.946 AUC (Disaster Model): AUC 0.946 at 30-yr horizon for area-level disaster prediction. Validated on 68,485 real FEMA disasters.
Industry Standard: 0.80 AUC is considered "good" - CLIMA's models exceed this benchmark.
Why it matters: Higher AUC = more reliable predictions = better lending decisions = fewer losses.

Credit Risk Metrics (PD, LGD, EL)

What they are: Three standard banking metrics that measure credit risk. Basel III and CECL require these for all loans (banks already calculate these). NEW: Regulators have issued guidance (not yet mandatory) expecting climate risk incorporation. CLIMA provides climate risk adjustments (PD uplift, LGD uplift, EL increase) that banks add to their existing calculations - we don't replace their PD/LGD/EL models, we enhance them with climate risk.

PD (Probability of Default)

The likelihood a borrower will default on their mortgage. CLIMA shows how climate risk increases this probability. Example: If base PD is 2% and climate adds +4%, the adjusted PD is 6%.

LGD (Loss Given Default)

The percentage of loan value lost when a borrower defaults. Disaster-damaged properties have lower resale values, so LGD is higher. Example: If a property is destroyed by wildfire, you might recover only 40% of the loan value (60% LGD).

EL (Expected Loss)

Total credit risk = PD × LGD × EAD (Exposure at Default). This is the dollar amount you expect to lose. Used for setting aside reserves and calculating capital requirements. Example: If EL increases by 451%, you need 5.5x more reserves.

Why These Metrics Matter

For Regulatory Compliance:

Basel III and CECL require PD, LGD, and EL calculations for all loans (banks already do this). NEW: OCC, FDIC, and Fed have issued guidance (not yet mandatory requirements) expecting banks to incorporate climate risk into credit risk assessments. CLIMA provides climate risk adjustments (PD uplift, LGD uplift, EL increase) that banks add to their existing PD/LGD/EL calculations, helping them meet regulatory expectations and prepare for future requirements.

For Risk Management:

These metrics tell you exactly how much climate risk costs. A 451% EL increase means you need 5.5x more reserves - critical information for capital planning.

Why CLIMA Wins

  • End-to-End: Scoring + recommendations + analytics in one platform
  • Credit Integration: Climate risk adjustments (PD/LGD/EL uplifts) built-in
  • Fast Implementation: 2-4 weeks vs. 6-12 months
  • Lower Cost: $50K-$150K vs. $2M+ for competitors
  • Validated Models: 0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC, 145K observations

ROI Comparison

CLIMA
Cost: $75K/year
Savings: $2.4M/year
35-50x ROI
Competitors
Cost: $2M+ (Year 1)
Savings: Similar
Negative ROI (Year 1)

What Makes CLIMA Different

  • Actionable Recommendations: Not just scores—actual lending terms
  • Mortgage-Specific: Built for lenders, not insurers
  • Regulatory Ready: Basel III, CECL compliance built-in
  • No Internal Build: Ready to use, no engineering team needed
  • Proven Results: 96% loss reduction validated (Freddie Mac 15M loans)

CLIMA Pricing

Simple, transparent pricing based on portfolio size

Community Banks

$50K-$100K
5K-50K mortgages
  • Full portfolio scoring
  • Lending recommendations
  • Regulatory reports
  • API access
MOST POPULAR

Regional Banks

$100K-$250K
50K-500K mortgages
  • Full portfolio scoring
  • Lending recommendations
  • Regulatory reports
  • API access
  • Priority support
  • Custom analytics

Large Banks

$250K-$750K
500K+ mortgages
  • Full portfolio scoring
  • Lending recommendations
  • Regulatory reports
  • API access
  • Dedicated support
  • Custom integrations
  • White-label options

All plans include: Production ML models (0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC), complete analytics, regulatory compliance, and ongoing support.

Complete Portfolio Solutions

Everything we deliver for your entire portfolio. Production-grade ML models, comprehensive analytics, and regulatory-ready reporting.

Portfolio Scoring

Score thousands of loans in minutes with our production ML models (0.842 default / 0.946+ disaster (0.993 at 30-yr) AUC).

  • Batch processing for entire portfolios
  • ZIP to County to State aggregation
  • Risk tier distribution analysis
  • High-risk watchlist generation

Geographic Diversification

Identify concentration risks and optimize geographic allocation.

  • Regional concentration analysis
  • State-level risk heatmaps
  • County aggregation metrics
  • Concentration limit recommendations

Stress Testing

Multiple NGFS scenarios and custom stress tests for regulatory compliance.

  • NGFS Hot House World 2050
  • NGFS Orderly/Disorderly Transition
  • Custom scenario builder
  • Baseline vs. stressed comparisons

Regulatory Capital

Calculate capital requirements aligned with Basel III and CECL standards.

  • Risk-weighted capital calculations
  • CECL provision estimates
  • Capital buffer requirements
  • Regulatory compliance reporting

Risk Concentration

Identify high-risk concentrations and exposure hotspots.

  • Top N risk concentrations
  • Herfindahl-Hirschman Index (HHI)
  • High-risk area identification
  • Exposure percentage calculations

Portfolio Optimization

Optimize risk/return with AI-driven allocation recommendations.

  • Optimal allocation by risk tier
  • Revenue-preserving strategies
  • Portfolio improvement suggestions
  • Risk-adjusted return analysis

Executive Reports

Regulatory-ready PDF and Excel reports for board presentations.

  • Executive summary reports
  • Portfolio risk distribution
  • Stress test results
  • Methodology documentation

Correlation Analysis

Understand how loans fail together in disaster scenarios.

  • Geographic correlation risks
  • Hazard concentration analysis
  • Portfolio diversification metrics
  • Systemic risk identification

Data Intake Engine

Seamless portfolio data upload with intelligent validation.

  • CSV/Excel file upload
  • Automatic field mapping
  • Data quality validation
  • Batch processing pipeline

What You'll Receive

Data Files

  • - Scored loan file (CSV/Parquet) with CLIMA scores
  • - Portfolio summary by risk tier
  • - County-level risk aggregation
  • - High-risk watchlist
  • - Geographic risk breakdown

Analytics & Reports

  • - Executive summary (PDF)
  • - Stress test results (multiple scenarios)
  • - Risk concentration analysis
  • - Portfolio optimization recommendations
  • - Methodology documentation
145,452
Observations Validated
68,485
Real Disasters Analyzed
~$47M
Annual Value ($10B example)
623x
ROI

Ready to Transform Your Risk Assessment?

Join leading financial institutions using CLIMA to reduce losses and improve portfolio performance.