Climate AI Nordics

Climate AI Nordics

Nordic network for collaborations around tackling climate change using AI and machine learning.

Severe decline in large farmland trees in India over the past decade

Authors: Martin Brandt, Dimitri Gominski, Florian Reiner, Ankit Kariryaa, Venkanna Babu Guthula, Philippe Ciais, Xiaoye Tong, Wenmin Zhang, Dhanapal Govindarajulu, Daniel Ortiz-Gonzalo & Rasmus Fensholt

Agroforestry practices that integrate multifunctional trees within agricultural landscapes can generate a range of socioecological benefits, while also serving as a natural climate solution due to their carbon sequestration potential. These trees form a vital part of India’s landscapes. This study focuses on how agroforestry systems in India have transformed over the last decade.

Nationwide mosaics for RapidEye for 2010 and 2011 (5 m resolution) and for PlanetScope (3 m resolution) for 2018, 2019, 2020, 2021 and 2022, along with a large corpus of tree labels for PlanetScope (about 130,000 labels) and RapidEye (about 100,000 labels) forms the basis of the study. Since RapidEye is unreliable in identifying small trees, the study focused on large and mature trees, which can be detected with high confidence (>0.8)) in these images. Deep learning models inspired by the object counting literature were used to detect trees via heatmap-based methods, transforming discrete point labels into continuous Gaussian kernels to enable smoother optimization.

If a tree crown center was visible in the early period (2010/2011) but absent in the later period (2018–2022), it was classified as having disappeared. The results show that 11 ± 2% of high-confidence trees mapped in 2010 or 2011 were no longer present in the 2018–2022 imagery. Several hotspot areas lost up to 50% of their large farmland trees, with as many as 22 trees per square kilometre disappearing. Focusing on the period from 2018 to 2022, 5.3 million trees (2.7 trees per square kilometre) observed in 2018/2019 were not detected in 2020–2022.

To understand the drivers of this change, they conducted interviews with villagers from Telangana, Haryana, Kerala, Maharastra, Andrah Pradesh, Uttar Pradesh, and Jammu and Kashmir. Alterations in cultivation practices, such as the establishment of new boreholes and expansion of paddy rice fields with the objective of boosting crop yields, were identified as an important factor in the rapid decline of agroforestry trees. While all interviewees reported that native trees within fields have become rare, several reported that block plantations have increased, highlighting the rapidly changing farming practices in India.

Published:

Categories:

You May Also Enjoy

Climate AI Nordics Newsletter, May 2025

Published:

Welcome to the Climate AI Nordics Newsletter, May 2025! Catch up on the Nordic Workshop on AI for Climate Change, major tree loss in India’s agroforestry, and upcoming events on conservation, causal inference, and digital emergency systems.

High-stakes decisions from low-quality data: AI decision-making for conservation

Event date:

Webinar with Lily Xu, Columbia University. Like many of society's grand challenges, biodiversity conservation requires effectively allocating and managing our limited resources in the face of imperfect information. My research develops data-driven AI decision-making methods to do so, overcoming the messy data ubiquitous in these settings. Here, I’ll present technical advances in machine learning, reinforcement learning, and causal inference, addressing research questions that emerged from on-the-ground challenges in wildlife conservation. I’ll also discuss bridging the gap from research and practice, with anti-poaching field tests in Cambodia, field visits in Belize and Uganda, and large-scale deployment with SMART conservation software.

Debiasing AI predictions for causal inference without fresh ground truth data

Event date:

Webinar with Markus Pettersson, Chalmers University of Technology. Machine learning models trained on Earth observation data, particularly satellite imagery, have recently shown impressive performance in predicting household-level wealth indices, potentially addressing chronic data scarcity in global development research. While these predictions exhibit strong predictive power, they inherently suffer from shrinkage toward the mean, resulting in attenuated estimates of causal treatment effects and thus limiting their utility in policy evaluations. Existing debiasing methods, such as Prediction-Powered Inference (PPI), require additional fresh ground-truth data at the downstream causal inference stage, severely restricting their applicability in data-poor environments. In this paper, we introduce and rigorously evaluate two novel correction methods—linear calibration correction and Tweedie's correction—that substantially reduce prediction bias without relying on newly collected labeled data. Our methods operate on out-of-sample predictions from pre-trained models, treating these models as black-box functions. Linear calibration corrects bias through a straightforward linear transformation derived from held-out calibration data, while Tweedie's correction leverages empirical Bayes principles to directly address shrinkage-induced biases by exploiting score functions derived from predicted outcomes. Through analytical exercises and experiments using Demographic and Health Survey (DHS) data, we demonstrate that both proposed methods outperform existing data-free approaches, can achieve significant reductions in attenuation bias and thus providing more accurate, actionable, and policy-relevant estimates. Our approach represents a generalizable, lightweight toolkit that enhances the reliability of causal inference when direct outcome measures are limited or unavailable.

Inaugural Nordic Workshop on AI for Climate Change sparks collaboration and looks to the future

Published:

The inaugural Nordic Workshop on AI for Climate Change in Gothenburg successfully united leading researchers, practitioners, and innovators, fostering lively interdisciplinary discussions and idea exchange on topics ranging from biodiversity monitoring to the environmental footprint of AI. Building on this enthusiastic and collaborative atmosphere, the community is now looking forward to future events, with the 2026 workshop already confirmed to be hosted by the University of Copenhagen and the Pioneer Centre for AI. This vital initiative continues to build a strong Nordic network dedicated to maximizing AI's potential in addressing the urgent climate crisis.