case studies.

AI Revolutionises Biodiversity Conservation and Monitoring

Revolutionising Biodiversity Tracking and Habitat Preservation with Deep Learning Powered Computer Vision

Revolutionising Biodiversity Tracking and Habitat Preservation with Deep Learning Powered Computer Vision

Nature enthusiasts, conservationists, and researchers alike recognise the importance of continuously monitoring biodiversity hotspots and habitats around the globe. Traditional surveying techniques, however, impose financial burdens and labour intensity, hindering widespread deployment.

Thankfully, the advent of computer vision and deep learning heralds cost-effective alternatives, enabling automatic, scalable, and objective ecological assessment. Today, we delve into the compelling advantages of marrying these sophisticated image recognition tools with terrestrial and marine conservation efforts.

Deep Learning Powers Accurate Ecological Health Diagnostics:

Deep learning models, inspired by biological neuronal architectures, possess extraordinary pattern recognition abilities. Consequently, these algorithms assimilate immense volumes of labeled photographic evidence, discerning subtle signs of flora and fauna amid complex backgrounds. Compared to manual photo tagging, computers readily distinguish cryptic organisms and nuanced ecological states, minimizing observer bias and reducing operational costs.

Real World Applications Abound:

1. Tree Cover Estimation: State-of-the-art convolutional neural networks quantify tropical deforestation and afforestation dynamics, complementing official statistics reported by national governments. Such independent verification strengthens confidence in internationally recognized carbon credits.

2. Coral Reef Vitality Indexing: Semi-supervised deep learning models gauge coral abundance, diversity, and bleaching severity within coastal waters. Timely warnings issued to local administrators galvanize protective measures, averting irreversible damage caused by pollution, sedimentation, and climate change.

3. Avian Occupancy Mapping: Object-detecting algorithms pinpoint avian rookeries and migratory routes, revealing previously unknown concentrations of threatened species. Detailed cartography illuminates priority regions for habitat restoration and protected zone establishment.

4. Aquatic Macrofaunal Census: Neural networks trained on underwater footage reliably count and catalogue plankton, crustaceans, molluscs, and fish inhabiting estuaries, lakes, rivers, and seas. Comprehensive inventories reveal anthropogenic stressors acting upon native fauna and invasive alien species establishing themselves.

Overcoming Barriers:

Implementing computer vision techniques requires substantial initial investments and technical proficiency. Partnering with academic institutions, government entities, and nonprofits pools intellectual capital and shares expenses, hastening the pace of ecological recovery. Similarly, sponsoring workshops and hackathons fuels grassroots enthusiasm, inspiring students, professionals, and retirees to volunteer their talents toward noble causes. Finally, advocacy drives legislation mandating private enterprise investment in sustainability programs, catalysing corporate citizenship and lasting environmental protections.


Combining computer vision and deep learning promises monumental strides toward understanding and combatting biodiversity losses and habitat degradation. As pioneers blaze trails in this nascent discipline, expect heightened appreciation for the planet’s breathtaking beauty, profound complexity, and astonishing resiliency. Together, let’s celebrate the dawn of a golden era in environmental protection, fueled by imagination, determination, and groundbreaking technology.


  1. Nazir, S. and Kaleem, M., 2021. Advances in image acquisition and processing technologies transforming animal ecological studies. Ecological Informatics, 61, p.101212.
  2. Klein, D.J., McKown, M.W. and Tershy, B.R., 2015, May. Deep learning for large-scale biodiversity monitoring. In Bloomberg Data for Good Exchange Conference (Vol. 10).
  3. Christin, S., Hervet, É. and Lecomte, N., 2019. Applications for deep learning in ecology. Methods in Ecology and Evolution, 10(10), pp.1632-1644.
  4. Beery, S.M., 2023. Where the Wild Things Are: Computer Vision for Global-Scale Biodiversity Monitoring. California Institute of Technology.
  5. Chisom, O.N., Biu, P.W., Umoh, A.A. and Obehioye, B., 2024. Reviewing the role of AI in environmental monitoring and conservation: A data-driven revolution for our planet.
  6. Lahoz-Monfort, J.J. and Magrath, M.J., 2021. A comprehensive overview of technologies for species and habitat monitoring and conservation. BioScience, 71(10), pp.1038-1062.

Harnessing AI and Machine Learning to Optimize Environmental Research Data Analysis

Harnessing AI and machine learning for optimising environmental research data analysis

Artificial intelligence (AI) and machine learning (ML) revolutionise environmental research by streamlining big data analysis, enabling accurate predictions, and fostering cross-domain collaboration. Key applications include:

1. Automated Classification: Supervised learning algorithms categorise features like land cover, species distributions, and water quality parameters using multispectral satellite imagery, airborne sensors, or ground-based observations.

2. Anomaly Detection: Detect irregularities, extremes, or abnormal behaviours in ecological systems, triggering timely intervention efforts.

3. Predictions: Anticipate future conditions, interactions, or consequences based on historical data, statistical correlations, or mechanistic simulations.

Benefits extend beyond single studies to broader collaborations:

1. Data Fusion: Harmonise disparate sensor networks, observational platforms, or citizen science contributions into coherent datasets ripe for comparison, validation, and metadata exchange.

2. Semantic Interoperability: Ensure consistent semantic encoding, retrieval, and aggregation of geospatial information assets using shared terminologies and ontologies.

3. Open Science: Foster transparency, reproducibility, and repeatability through version control systems, open-access publishing, and crowdsource competitions.

However, certain challenges persist:

1. Interpretability: Craft transparent, explainable models for domain experts.

2. Data Scarcity: Handle imbalanced, sparse, or missing observations affecting training robustness and transferability.

3. Ethical Concerns: Safeguard vulnerable populations and natural resources from potential misuse or exploitation.

Responsibly wielding AI and ML empowers environmental researchers to tackle pressing challenges and drive progress toward sustainable development goals.



Foreign Policy New Realities

2018-present. Complete web design/development and support.

Project Overview

EU foreign policy experiences unprecedented turbulences that put key achievements of the European integration project at risk. Externally, the EU’s global environment is characterized by the reconfiguration of power, growing divisions, and the contestation of the established liberal order.

Simultaneously, the EU’s neighborhood is increasingly conflict-prone and unstable, triggering migration flows and the proliferation of illiberal values. 

Domestically, the EU faces severe internal conflicts, marked by austerity, Brexit, growing nationalism, populism, and new protectionism.

COST Action ENTER aims to improve our understanding of central properties of EU foreign policy in light of these new realities, focusing on perceptions, communication, contestation.

In today’s world, the success of EU foreign policy depends on the EU’s ability to instantaneously respond to stimuli and pressures originating from both the international and the intra-EU levels.

The Action ENTER aims to improve our understanding of central properties of EU foreign policy in light of these new realities, focusing on perceptions, communication, contestation.

In today’s world, the success of EU foreign policy depends on the EU’s ability to instantaneously respond to stimuli and pressures originating from both the international and the intra-EU levels.

Linking internal and external policy dynamics, the Action has a strong potential for breakthrough scientific developments.



2014-present. Complete web design/development and support.

INOGOV is a network of research excellence dedicated to understanding the sources, patterns, and effects of policy and governance innovations for climate change.

Set against scientific predictions, current international responses to climate change are widely perceived to be inadequate.

There is a growing perception, that many mitigations and adaptation measures have been taken outside the international regime.

In this sense, governance has become considerably more polycentric, with pockets of dynamism especially evident at the national and subnational levels, but also in the so-called transnational sphere.

However, there is far less agreement on if and how these innovations can be scaled up, if and indeed how they should be coordinated, and where the necessary leadership to achieve this might originate.

"atoms and bits are excellent Web Developers who have developed a number of websites to an extremely high standard. Their attention to detail, both in terms of client/customer service as well as site content is exceptional. They are also a pleasure to work with."
Adapt Lock-in Project

Adapt Lock-in

2020-present. Complete web design/development and support.

Project overview

Preparing for and coping with the accelerating impacts of climate change requires adaptation in a wide range of policy areas. Yet, despite increasing calls for action, policy change to allow this is often slow.

A range of counteracting forces and barriers can make it difficult to embed adaptation objectives into important policies and move them away from ‘business-as-usual. 

However, deeper dynamics are also at play, where self-reinforcing mechanisms, feedbacks, and path dependencies interact across different Spatio-temporal scales and coalesce to establish policy lock-ins.

The aim of this interdisciplinary project is to uncover these lock-in dynamics and examine the extent to which they account for varying levels of climate change adaptation in Germany, the Netherlands, and the U.K. (England).

Project duration

June 2019 – May 2022

Professor Mike Hulme

Professor Mike Hulme

2007-present. Complete blog development and continuous support.

Mike is a Professor of Human Geography in the Department of Geography at the University of Cambridge and a Fellow of Pembroke College, where he is the Director of Studies for Geography.

His work explores the idea of climate change using historical, cultural, and scientific analyses.

He seeks to illuminate the numerous ways in which climate change is deployed in public and political discourse and to this end, he is currently finishing a book manuscript on ‘The Idea of Climate Change‘ for the Routledge Key Ideas in Geography book series, due for publication in June 2021.

"atoms and bits respond promptly and accurately whenever there is an issue with the website. Whether a technical issue or a legal or security one, Javier is on top of them all and knows how to fix things."


2016-present. Complete web design/development and support.

Project overview

The UKESM project is charged with developing, and applying UK Earth system models (UKESM) and providing user support for these models to the wider NERC/academic community.

UKESM project aims are:

  • To apply UKESM1 to investigate a range of Earth system phenomena and their sensitivity to future anthropogenic forcing.

  • To scientifically evaluate and document the performance of UKESM1.

  • To analyze and document UKESM1 future projections.

  • To provide science-based guidance on future Earth system change.

  • To provide the tools and user support for UK Earth system modeling research.

  • To initiate the development of a future UKESM2 model.


Project duration

2017 – 2021

Tyndall Centre

Tyndall Centre for Climate Change Research

2014-present. Complete web design/development of two previous website versions and continuous support.

We provide evidence to inform society’s transition to a sustainable low-carbon and climate-resilient future.

The Tyndall Centre is a partnership of universities bringing together researchers from the social and natural sciences and engineering to develop sustainable responses to climate change. We work with leaders from the public and private sectors to promote informed decisions on mitigating and adapting to climate change.

“Above all, we undertake robust and independent research to identify the challenges and opportunities presented by climate change and inform open and transparent decisions that best serve society”.

About the Tyndall Centre: The Tyndall Centre was founded in 2000 to conduct cutting edge, interdisciplinary research, and provide a conduit between scientists and policymakers. With nearly 200 members ranging from PhD researchers to Professors, the Tyndall Centre represents a substantial body of the UK’s climate change expertise from across the scientific, engineering, social science and economic communities.

The Tyndall Centre has since 2000 significantly advanced the fundamental analysis of emission reduction from all major energy sectors, the understanding of climate impacts, risks, and adaptation options, the public perceptions of climate change, and the governance of climate negotiations and policymaking. From 2000 to 2010 the Tyndall Centre was core-funded to a total of £19m by the Natural Environment Research Council, the Engineering and Physical Science Research Council, and the Economic and Social Research Council. In the years since core funding came to end, Tyndall partners receive several million per year in project funding, including from public and private consultations on how to respond to climate change.