# Earthquake fragility curves for building structures

**Example**: Fragility functions defining the probability of reaching or exceeding different damage states for buildings under earthquake ground shaking.

## Step-by-step guidance

### 1. Dataset-level metadata

Select the following values when describing your dataset:

- **Risk data type**: `vulnerability`
- **Title**: "Earthquake fragility curves for \[building types/region\]"
- **Description**: Brief description of the fragility functions, building typologies covered, and derivation method
- **Publisher**: Organization or research institution that developed the fragility curves
- **License**: Appropriate license

### 2. Resources

Add resources for your fragility function files:

- **Format**: `csv`, `json`, or `xlsx`
- **Spatial resolution**: Not applicable (functions are typically building-type specific)
- **Coordinate reference system**: Not applicable

### 3. Vulnerability metadata

Under the Vulnerability section:

#### Category

- **Category**: `fragility`

#### Hazard type and intensity

- **Hazard type**: `earthquake`
- **Hazard primary type**: `earthquake`
- **Processes**: `ground_shaking`
- **Intensity measure**: Select appropriate measure:
  - `PGA:g` (peak ground acceleration in g)
  - `SA(0.3):g` (spectral acceleration at 0.3s)
  - `SA(1.0):g` (spectral acceleration at 1.0s)
  - `PGV:m/s` (peak ground velocity)

#### Vulnerability functions

For each building typology, create a function entry:

**Function specification:**

- **ID**: Unique identifier (e.g., `F1_RC_lowrise`)
- **Relationship**: `fragility`
- **Function type**: `lognormal`, `beta`, `empirical`, or `hybrid`

**Damage scale:**

- **Type**: `discrete` or `continuous`
- **Damage states**: List damage levels (e.g., `slight`, `moderate`, `extensive`, `complete`)
- **Number of damage states**: Count (e.g., 4)

**Parameters** (for lognormal):

- **Median** (μ): Intensity at 50% probability for each damage state
- **Beta** (β): Standard deviation of logarithm

#### Asset taxonomy

Link fragility curves to building types:

- **Taxonomy source**: `GED4ALL`, `HAZUS`, `PAGER`, `EMS-98`, or custom
- **Taxonomy code**: Specific code for each building type (e.g., `CR/LWAL+DNO/H:1,2`)

#### Cost information

If economic loss is included:

- **Cost dimension**: `structure`, `content`, `business_interruption`
- **Cost unit**: Currency code (e.g., `USD`)

### 4. Spatial coverage

Define applicability:

- **Scale**: `global`, `national`, or `regional`
- **Countries**: Select applicable ISO 3166-1 alpha-3 country codes if region-specific
- **Description**: Note if curves are calibrated for specific seismic regions

## Example data structure

Your fragility curve file should include:

**For each building type and damage state:**

- Building taxonomy code
- Damage state name
- Hazard intensity measure (e.g., PGA)
- Distribution type (lognormal, beta, etc.)
- Median value (50th percentile intensity)
- Dispersion (beta/standard deviation)
- Intensity units

**Example CSV structure:**

```
Building_Type,Damage_State,Intensity_Measure,Distribution,Median,Beta
RC_Lowrise_1-3stories,Slight,PGA:g,lognormal,0.15,0.45
RC_Lowrise_1-3stories,Moderate,PGA:g,lognormal,0.30,0.45
RC_Lowrise_1-3stories,Extensive,PGA:g,lognormal,0.60,0.50
RC_Lowrise_1-3stories,Complete,PGA:g,lognormal,1.20,0.55
```

## Key considerations

- Clearly document the derivation method (empirical, analytical, expert judgment, hybrid)
- Specify building taxonomy codes consistently
- Reference the intensity measure definition and units
- Include uncertainty information (beta values for lognormal distributions)
- Document the damage state definitions used
- Note any regional calibration or adjustments
- Include references to source publications or studies