Climate Vegetation Dynamics

This project presents a machine learning and web-based platform for monitoring, forecasting, and supporting adaptive responses to changes in vegetation and wind patterns driven by climate change.

Problem Statement

Climate change is altering vegetation patterns, rainfall, and wind dynamics, creating both environmental and societal challenges. This project addresses these shifts by building an AI/ML-powered platform that integrates real-time sensor simulation and forecast visualization to support adaptation decisions.

Project Objectives

• Predict NDVI (vegetation index) and wind conditions using machine learning models
• Simulate real-time environmental sensor data
• Provide actionable adaptation recommendations based on forecasted conditions
• Deliver insights through a web dashboard interface built with Streamlit

Key Components

Machine Learning Forecasting

• NDVI Prediction: Multi-horizon forecasting (t+1, t+7, t+16, t+30) using XGBoost regression
• Wind Forecasting: Wind speed and wind direction forecasts (t+1, t+3, t+7)
• Feature Engineering: Includes lag features, rolling statistics, wind vector transformation, and seasonal features
• Model Evaluation: RMSE, R² metrics for all forecast horizons

Adaptation Decision Engine

A rule-based system generates dynamic adaptation advice based on:

• Low NDVI levels
• Insufficient rainfall
• High wind speeds

Examples:

• Recommend irrigation under vegetation stress
• Warn against pesticide application during high winds

Streamlit Dashboard

• Real-time Sensor Simulation: Generates randomized but realistic data streams
• Forecast Visualizations: Line plots comparing actual vs predicted NDVI and wind
• Adaptation Alerts: Instant decision support messages based on conditions

Repository Structure

climate_vegetation_dynamics/ ├── climate_model.py # Complete ML forecasting pipeline
├── streamlit_app/
│ ├── dashboard.py # Streamlit frontend interface
│ └── assets/ # Forecast visuals, plots
├── outputs/ # Forecast results (CSVs, plots, metrics)
├── data/ # Sample input datasets
├── requirements.txt # Python dependencies
├── .gitignore # Files/folders excluded from version control
├── LICENSE # MIT License (optional)
└── README.md # Project overview and documentation

Getting Started

Install Dependencies

pip install -r requirements.txt streamlit run streamlit_app/dashboard.py python climate_model.py

Dependencies

• Python 3.8+
• pandas
• numpy
• matplotlib
• seaborn
• scikit-learn
• xgboost
• streamlit

Demo