A satellite foundation model for improved wealth monitoring

2026-04-25

AI Summary

Traditional poverty statistics rely on costly, infrequent household surveys that struggle to capture localized economic dynamics. This work proposes Tempov, a satellite foundation model that leverages self-supervised pretraining on three million pairs of bitemporal Landsat images and parameter-efficient fine-tuning to enable high-resolution monitoring of wealth dynamics under extremely sparse survey labels. Tempov is the first method capable of zero-shot nowcasting, retrospective prediction, and change tracking across a decade, substantially reducing reliance on ground-based data. Using only 10% of available survey samples, it achieves an out-of-sample R² of 0.63 (r = 0.68) across Africa, producing decadal high-resolution maps of wealth change that reveal pronounced intra- and inter-country economic disparities.

Technology Category

• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Computer Vision: Remote Sensing / Geospatial AI
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: The sharing economy
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Poverty statistics guide social policy, but in many low- and middle-income countries, censuses and household surveys that collect these data are costly, infrequent, quickly outdated, and sometimes error-prone. Satellite imagery offers global coverage and the possibility of predicting economic livelihoods at scale, yet existing approaches to predicting livelihoods with imagery or other non-traditional data often fail to reliably identify local-level variation and, as we show, degrade under temporal shift. Here we introduce Tempov, a satellite foundation model pretrained by self-supervision on three million bi-temporal Landsat pairs and adapted with parameter-efficient fine-tuning to sparse survey labels. The model enables large-scale, high-resolution wealth mapping and dynamic measurement, including zero-shot nowcasting up to a decade after observed labels, retrospective hindcasting, and decadal change tracking, while outperforming existing neural network and geospatial foundation-model baselines. In low-label regimes, Tempov achieves competitive accuracy with only 10% of survey samples, indicating substantially reduced dependence on expensive label collection. The model further generalizes across populous countries within and outside Africa, and scales to a unified Africa-wide model with strong continent-level performance ($R^2=0.63$, $r^2=0.68$), from which we generate high-resolution decadal maps of wealth and wealth changes for the African continent. Analysis of these maps shows large variation in recent economic performance both within and across countries. Our open-source approach provides a pathway to timely, scalable, low-cost monitoring of wealth and poverty from routinely collected satellite data.

Problem

• poverty monitoring
• satellite imagery
• economic livelihoods
• temporal shift
• wealth mapping

Innovation

• satellite foundation model
• self-supervised pretraining
• parameter-efficient fine-tuning
• wealth mapping
• temporal generalization