The core challenge in planning for forest resilience across Europe is understanding which regions today will share similar climatic conditions with projected future climates. I was contracted to take over an existing, but defunct, web application designed to address this. The original tool had stopped functioning, its UI was outdated, and its backend calculations were slow due to a lack of optimisation.

Phase 1: Performance Engineering and Code Modernisation

My first task was technical triage and optimisation. The existing application's backend was not optimised, resulting in long calculation times. The primary goal was algorithmic efficiency, to support an interactive web experience.

Strategic Performance Optimisation with NumPy

I completely re-engineered the core computational logic. I implemented a new Python engine using NumPy for high-dimensional array processing. I used vectorisation to handle the matching equations (based on mean monthly temperature, diurnal temperature range, and precipitation, as described in Broadmeadow et al. (2005)).

This re-engineering transformed the tool, achieving much faster result delivery and making the entire application viable for production use.

Phase 2: Full-Stack Web Application Redevelopment

With the core performance issues solved, I collaborated with Forest Research to deliver a complete, modern replacement. In this phase, I focused on designing and deploying a robust, user-friendly application, now hosted by Forest Research at climatematch.org.uk.

API Wrapper and Modern UI/UX

  • Backend Service: I developed a new API wrapper around the optimised Python engine using Flask. This framework efficiently handles client requests, executes the fast matching calculations, and serves structured data to the client.
  • Frontend and UI/UX: I designed and developed a completely new user interface using responsive design patterns, with React, Leaflet and Plotly. This not only modernised the look of the application but ensured it is equally easy to use on smartphones and desktop computers.

Delivering Actionable Insights for Forest Resilience

The revitalised tool provides forestry practitioners and policy makers with an intuitive way to visualise potential future climate (using the Representative Concentration Pathway (RCP) 8.5 scenario). It helps them select seeds that will maintain forest resilience and plan for potential future pest species. It also includes a frequency distribution to aid the user's interpretation of matched climatic conditions.

I developed and modernised this tool as part of my ongoing work at Forest Research, with funding support from both B4EST and the Forestry Commission.

See the Project in Action

Explore the live application and the underlying code that powers this tool: