User Guide

Learn how to use the Philippines Landslide Forecasting System

Getting Started

Welcome to the Philippines Landslide Forecasting System! This guide will help you navigate and use all features effectively.

What is this system?

Our system uses advanced AI and machine learning to predict landslide risks across the Philippines. It analyzes multiple factors including:

  • Rainfall data - Both cumulative and intensity measurements
  • Soil moisture - Ground saturation levels
  • Terrain analysis - Slope angles
  • Soil composition - Different soil types

Using the Map

1. Selecting a Location

You can select a location in two ways:

  • Search: Type a city name in the search box (e.g.,"Manila", "Cebu", "Davao")
  • Click: Click anywhere on the map to select that location
Tip: The map covers the entire Philippines. You can zoom in/out and pan around to explore different areas.
2. Understanding Map Markers

When a location is selected, key geospatial data is displayed:

  • Location: This displays the location name
  • Slope: This displays the slope angle of the location
  • Soil: This displays the soil composition of the location

Making Predictions

Step-by-Step Process
1
Select Location

Search or click on the map

2
Choose Date

Select forecast date (up to 5 days ahead)

3
Select Period

Choose from 3hrs to 5 days

4
Get Prediction

Click "Predict Landslide "

Available Time Periods:
  • 3 Hours - Urgent / Immediate Risk
    Use for immediate safety decisions.
  • 6 Hours - Extended immediate risk
    Use for advanced warnings to specific areas.
  • 12 Hours - Half-day forecast
    Use for making decisions that affect a whole day.
  • 1 Day - Full day forecast
    Use for strategic planning and general awareness.

The Use Of Time Period

Step-by-Step Process

1

The "Snapshot" Prediction

First, our AI system takes a "snapshot" of all current conditions (rainfall, soil, slope) to calculate the landslide risk for one specific moment in time.

2

Checking Every Hour

To forecast over a time range (e.g., 6 hours), the system runs a snapshot prediction for every single hour in that period.

3

Identifying the Moment of Highest Confidence

At each hourly check, our model provides two key pieces of information:

  • A Prediction: Either "Landslide" or "No Landslide".
  • A Confidence Score: A percentage telling us how certain the model is about its prediction.

Our system is designed for safety, so it reviews all the hourly checks and shows you the single highest "Landslide" confidence score it found. This ensures you are always alerted to the moment of peak danger.

Understanding Results

Prediction Result Explained
Landslide Predicted: A high probability of a landslide event. Take immediate precautions.
No Landslide Predicted: Conditions are currently stable with low probability.
Contributing Factors
Factor Description Impact
Rainfall Cumulative rainfall (mm) and intensity. High values significantly increase risk.
Soil Moisture Water content in the soil (%). Saturated soil is less stable.
Slope Angle of the terrain in degrees. Steeper slopes have higher risk.
Soil Type Composition and characteristics of the soil. Clay-heavy soils are more susceptible.

Dashboard Features

The dashboard provides comprehensive system monitoring and historical analysis.

Key Metrics
  • Historical Occurrences: Yearly/monthly landslide counts.
  • Geospatial Map: Location of past landslide events.
  • Regional Hotspots: Regions with the highest frequency of landslides.
Charts & Analysis
  • Soil Vulnerability: Count of landslides per soil type.
  • Slope Analysis: Relation of slope angle to landslide occurrences.

Frequently Asked Questions

Our system maintains a 90.6% accuracy rate based on historical validation. However, predictions should be used alongside other safety measures and professional judgment.

Contact local authorities, avoid the high-risk area if possible, and follow official emergency protocols. Consider evacuation if recommended by local disaster management agencies.

Weather data is updated every 3 hours, terrain data is static but highly accurate, and the ML model is retrained weekly with new data to maintain accuracy.

While our system is highly accurate, it should be used as a supplementary tool alongside official meteorological and geological services. Always consult with local authorities for official emergency decisions.