CLIMATE CHANGE ANALYSIS
USING MACHINE LEARNING
TELUGU VENKATA TEJASWINI
160621733184
INTRODUCTION
• Machine Learning is a fascinating subset of Artificial Intelligence
(AI). It revolves around creating algorithms that empower
computers to learn autonomously from data and past experiences,
without explicit programming.
• Unlike traditional rule-based programming, ML models adapt and
improve by analyzing data
• It’s like teaching a computer to recognize patterns and make
decisions based on examples.
CLIMATE CHANGE ANALYSIS
As the climate crisis intensifies, AI is emerging as a tool for climate action,
including modeling, prediction and optimization.
Machine learning's some of complex effects on climate change:
Information Forecasting and
1. gathering, such as 2. 3 Automated
simulation, such as decision-making,
analyzing satellite using historical data such as optimizing
and aerial imagery and imagery to buildings' heating
to track predict extreme and cooling systems
deforestation. weather events based on weather
conditions and
LITERATURE SURVEY
• ML in Climate Analysis:
• ML enhances accuracy in climate predictions compared to traditional models.
• Applications: Regional climate downscaling, rainfall forecasting, and health
impact predictions.
• Long-Term Climate Prediction:
• ML predicts long-term climate responses using short-term simulation data.
• Outperforms traditional methods in capturing regional and short-lived forcing
effects.
• Global Warming Predictions:
• CO2 identified as the primary driver of warming.
PROPOSED SYSTEM
We construct the mapping between short-term temperature response and
long-term temperature response using Ridge regression and Gaussian
Process Regression (GPR). These include Random Forest.
•Ridge Regression tweaks the standard linear regression by adding a
penalty term to the cost function. This penalty discourages large
coefficients, making the model less prone to overfitting.
•GPR: When you don’t know the exact relationship between variables but
want to make predictions we use GPR. It is great for continuous data and
handling uncertainty.
ADVANTAGES
•Helps track gradual changes over time.
•Evaluates climate impacts across various
sectors.
•Gives better performance especially in
regions with poor station data availability.
CONSOLIDATION OF PAPERS
ML for Regional Climate Change Projections
• ML predicts long-term temperature responses using short-term data,
reducing computational costs and enhancing climate projections.
Tropical Storm Identification with ML
• Transfer learning applied to classify tropical storms with fewer input
data (2–10x less) and more intensity classes.
Greenhouse Gases and Climate Prediction
• ML models validate predictions and address challenges like the faint
young Sun paradox.
CONCLUSION AND FUTURE SCOPE
• While there are challenges associated with climate change analysis
using surrogate models, these tools offer a promising pathway towards
enhancing our understanding of climate dynamics.
• By leveraging simplified representations of complex systems, surrogate
models provide valuable insights that can inform mitigation strategies
and adaptation efforts.
•Temperature changes can be predicted which comes from greenhouse gas
emissions.
•It can analyze the melting of snow and glaciers due to increase in
temperature.
•Changing in precipitation can be predicted.
THANK
YOU
