Abstract
<title>Abstract</title> <p>We introduce the Eco-Emotion Image ATLAS (E2IATLAS), a normative database of 3,161 nature photographs rated by 2,092 participants from 52 countries on six affective dimensions: valence, arousal, awe, compassion, connectedness, and gratitude. Images were assembled from open-access affective-science databases and photographic repositories, spanning eight content categories from pristine landscapes to human-impacted environments, with about 50 raters per image sampled across all five major climate zones. Factor analysis and network modelling converged on a three-factor structure: a dominant Positive Transcendence dimension uniting valence, awe, connectedness, and gratitude; a largely orthogonal Activation factor defined by arousal; and a partially independent Compassion factor. Cluster analysis identified four image profiles, including a Distress-Compassion profile — negative valence with elevated compassion — concentrated in environmental-issue imagery and invisible in valence-arousal space. Semantic content predicted affective responding substantially better than low-level visual features, with the built-infrastructure ratio the dominant suppressor of self-transcendent responding. Hedonic valence was near-universal across cultures, whereas self-transcendent dimensions showed moderate regional variation and a continuous latitude gradient. E2IATLAS is a shared platform for affective neuroscience, environmental psychology, and conservation-communication research.</p>