Photo by Billur Bektaş
Overview
Ecological forecasting aims to anticipate how populations and communities will change over time. Despite rapid growth in ecological data, forecasting ecological dynamics remains challenging due to delayed responses, transient dynamics, and context-dependent processes that limit predictability.
In this project, I work on improving ecological forecasting by developing, evaluating, and synthesizing near-term forecasts of ecological change. My contributions focus on linking ecological theory, data integration, and forecast evaluation to identify what aspects of ecological change are predictable, over which time horizons, and under which conditions.
This work spans multiple coordinated efforts, including European forecasting infrastructure projects, applied forecasting challenges, and the development of shared forecasting resources.
Research questions
- Which ecological variables and processes can be predicted reliably in the near term?
- How do delayed and transient ecological responses affect forecast performance?
- How does forecast skill vary across systems, models, and environmental contexts?
- How can forecasting workflows be standardized, evaluated, and reused across studies?
What I study in this project
Forecast development and evaluation
I develop near-term ecological forecasts and evaluate them against new observations, quantifying forecast accuracy, uncertainty, and sources of error.
Predictability and temporal dynamics
I examine how response lags and transient dynamics influence the ability to anticipate ecological change, and how predictability varies across systems and timescales.
Data integration and synthesis
I work on integrating heterogeneous ecological data into forecasting workflows and on synthesizing forecasts across models and datasets.
Capacity building and coordination
A key component of this project is supporting training and coordination in ecological forecasting, including shared databases, standards, and educational resources.
Subprojects and collaborative framework
PREDICT (OSCARS)
Within the OSCARS-funded PREDICT project, led by Maria Paniw (CSIC), I contribute to the development of an open and interoperable cyberinfrastructure for ecological forecasting, coordinated through LifeWatch ERIC. I work within a multidisciplinary team including Owen Petchey, Emily Simmonds, Thomas F. Johnson, Gerbrand Koren, Cara Gallagher, Melina Kourantidou, Patricia Singh, İstem Fer, Iestyn Woolway, and Marco Baity Jesi, among others. My contributions focus on developing a temporal and spatial forecasting challenge through fully open source plant functional traits data in European Alps.
Doñana vegetation forecasting challenge
I contribute to forecasting vegetation dynamics in the Doñana National Park through the Doñana Digital Twin platform led by Maria Paniw (CSIC). This work focuses on near-term forecasts of vegetation change and on evaluating how environmental drivers and model structure affect predictive performance. The first version of this infrastructure is now available at Doñana Digital Twin.
European ecological forecasting database
As part of the European Ecological Forecasting Initiative (EEFI) Steering Committee, I contribute to building a comprehensive database of ecological forecasts in Europe. The database documents ongoing and completed forecasting efforts, including forecast targets, data sources, modeling approaches, and evaluation practices. Its aim is to map the current landscape of ecological forecasting in Europe and to identify patterns, gaps, and opportunities for advancing predictability across systems. Explore the database at EEFI Forecasting Database.