Jivesh Dixit - Projects

1. Operationalizing the Madden Julian Oscillation (MJO) forecast at NCMRWF. (Sep 2024 - Oct 2024)

Association: National Centre For Medium Range Weather Forecasting (Ncmrwf)

Details:

MJO is a very important phenomenon of rainfall variability in tropics at S2S scales. Prediction of phase of MJO, i.e. zonal location of the wave or pulse of cloud and rainfall that moves eastward near the equator, is very important for short to medium range rainfall prediction. I operationalized the MJO prediction at S2S timescales using the output of the Coupled NCMRWF Unified model (CNCUM) forecast. Our S2S forecast including MJO forecast is shared across various Met agencies.

https://www.ncmrwf.gov.in/product-main-mjo-erf

https://github.com/JiveshDixit/Compute-MJO

2. Creating a Simple Catastrophe model from scratch using building data from OpenStreetMap and cyclone data using ERA5. (Sep 2024 - Present)

Association: Indian Institute of Technology, Delhi

Details:

In this project I have created a catastrophe model using commercial and residential building data from OpenStreetMap. The data for the asset value of the building is not available with me, so I have just assumed some numeric values in USD for asset in this model. I have chosen Hurricane Harvey as the hazard for this model. To calculate hazardous meteorological conditions from this cyclone I have use winds, wind gust and mslp data from ERA5. This model very closely calculated the landfall date and time along with the amplitude of meteorological variables as confirmed with https://earth.nullschool.net/ . I have chosen Harvey as it is considered wettest hurricane with losses totalling to $158.8 billion (https://www.nesdis.noaa.gov/news/hurricane-harvey-look-back-seven-years-later).

https://www.jiveshdixit.in/blogs/simple-catastrophe-model-risk-assessment-from-a-from-cyclone-case

3. Prediction of Pacific Decadal Oscillation (PDO) index using machine learning (Ongoing) (Sep 2024 - Present)

Association: Indian Institute of Technology, Delhi

Details:

PDO is a very non-linear phenomenon consisting of multiple timescales. Hosted in the Pacific Ocean, PDO perturbs global climate at multi-year to decadal timescales through teleconnections. Prediction of the PDO index at multi-year timescales is quite desirable by many industries, such as the energy sector, infrastructure sector, insurance and reinsurance sector, agriculture sector, and many more.

I used a machine learning algorithm to predict the PDO index using predictors such as sea-surface salinity, sea-surface temperature, mean sea-level pressure, and geopotential height. I used a Fourier-CNN-GRU-based model in this project.

The validation data gives a very good skill in predicting the PDO index by capturing variability in the PDO index with around 98% accuracy.

4. xIndices: A xarray based module for computing SST trends and SST variability modes and other atmospheric variability modes

Association: Indian Institute of Technology, Delhi

Details:

In this project I have created a Python based API to help the scientific community interested in Climate variability indices and patterns such as The El Niño-Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), North Atlantic Oscillation (NAO), Atlantic Multi-decadal Oscillation (AMO) and other variability modes in user defined regions along with variance fraction explained by these variability modes.

Capability of this API is not limited to computation of variability modes. It also allows use to preprocess the data for computation such as reordering longitude from -180_180 to 0_360, renaming the names of dimensions for smooth uses across the functions, regridding of curvilinear or rectilinear grid to rectilinear grid of resolution defined by the user.

It also has post-processing capabilities such as visualizing the indices and patterns without much effort. I am committed to add more variability modes and other capabilities to the software.

https://xindices.readthedocs.io/en/latest/

https://www.jiveshdixit.in/blogs/xindices-a-xarray-based-python-package-for-climate-variability-modes

Page updated

Google Sites

Report abuse

Email: jiveshdixit@cas.iitd.ac.in