Postdoctoral Researcher

  • Location: Preferably at the Bodeker Scientific offices in Alexandra. If that is not possible then in Wellington to be close to our MetService collaborators. Otherwise, elsewhere in New Zealand, subject to discussion

  • Term: 18 months. Starting as soon as possible

  • Remuneration: NZ$75,000/year

About us

Bodeker Scientific is an independent research organisation recognised domestically and internationally for its expertise in atmospheric and climate research. The company operates as a group of collaborating researchers who work on shared research contracts under a commonly agreed company constitution. We focus on conducting high-quality research within the time and financial constraints imposed by our research contracts. To achieve this goal, decision-making within the company is inclusive with an emphasis on personal responsibility.

Bodeker Scientific is based in Alexandra, Central Otago, known for its active lifestyle which provides excellent opportunities for outdoor pursuits including mountain biking, cycling, kayaking and water sports, skiing, rock climbing, etc. It is in a central location, with easy access to popular locations such as Queenstown and Wanaka.

About the project

Bodeker Scientific was recently successful in obtaining $999,880 in MBIE Endeavour funding for a project titled Using artificial intelligence to improve weather forecasts. This project is being conducted in close collaboration with the Meteorological Service of New Zealand (MetService) and with various other domestic and international partner organisations.

Through this project we plan to apply artificial intelligence methods to develop a new way of generating weather forecasts, thereby producing high-resolution forecasts at a fraction of current costs. A neural network (NN) will be trained to learn how to generate weather at hyperlocal scales (several 100m) given data from a lower resolution numerical weather prediction (NWP) model - in this case the Weather Research and Forecasting (WRF) model used by MetService. While the initial training may be computationally expensive, once trained, the NN can be applied to any NWP forecast to fill in the missing detail inside each grid-cell, at negligible cost. This cost reduction means that we can generate higher resolution forecasts than are currently available, and process many more forecasts to produce probabilistic risk assessments of rare but highly damaging events. If successful, our fused-NN-NWP model will be capable of delivering new hyperlocal weather forecasts, enhancing the information provided to key users such as emergency managers and other decision-makers who rely on high quality NWP to save lives and protect property, and manage risks to minimise losses. The need for such forecasts will only increase as the frequency and severity of extreme weather events increase under climate change.

About the role

The post-doc will be working as part of a larger team that, collectively, has the skills and expertise required to conduct the proposed research. The post-doc will therefore receive full support to further develop their own skill-set while working on all core aspects of the project. Their responsibilities will include:

  • Participating in designing the architecture of the NN

  • Coding the NN

  • Training the NN on high resolution fields of core weather variables

  • Investigating the feasibility of incorporating the NN into the operational WRF process

This is very much a research project with the opportunity to publish the results in the international peer-reviewed literature; this will be an opportunity to grow your CV in the burgeoning research area of AI applications in weather and climate. If you have any questions regarding the scope of the role and/or the responsibilities associated with the role, please contact Dr Greg Bodeker (

Essential skills/attributes

  • PhD in climate science, computer science, information and computer engineering, or equivalent.

  • Some working knowledge of artificial intelligence with a focus on machine-learning/deep-learning.

  • Experience in using numerical models.

  • Adept in using Linux.

  • Adept in Python programming.

  • Well-developed written and verbal communication skills with some first author publications in the international peer-reviewed literature.

  • Self-motivated and works well in a team.

Preferred skills

  • Demonstration of deep-learning capability through published papers.

  • Experience in the use of NWP models and in particular the WRF model.

  • Fortran programming experience.

  • Experience in computing in the cloud, and in particular use of Amazon Web Services (AWS).

How to apply

Please provide a cover letter, a CV which directly addresses your strengths in each of the areas listed above, a pointer to any GitHub repositories of your code (if available), and the names and contact details of at least three referees. If you are short-listed, you will be interviewed, either in person or via Zoom, for this position. Applications must be emailed to Annabel Yip ( by the deadline of noon 31 January 2023.