Stereoscopic remote sensing observation for mountains: Advancing monitoring, modeling, and management

Mountains cover approximately 24% of the Earth’s land surface, providing crucial sources of the planet’s fresh water and supporting vital ecosystem services. However, under the pressures of climate change and human activities, mountain ecosystems are rapidly transforming, acting as sentinels of global change.1 With the relentless advancement of satellite constellations, unmanned aerial vehicles (UAVs), and ground observation networks, we are entering a new era of stereoscopic Earth observation. Stereoscopic remote sensing combines observations from multi-altitude platforms and technologies like LiDAR and SAR to capture Earth’s information across various layers, altitudes, and depths. This approach enables the acquisition of multi-modal, multi-resolution, multi-angle, multi-spectral, and multi-temporal stereoscopic observation data, promoting a comprehensive understanding of the Earth’s surface and its dynamic processes. However, given the inherently complex three-dimensional (3D) nature of mountains, there is a pressing need for tailored observation methods and specific observation goals to effectively address the distinct challenges posed by these environments.

Here, we delve into the connotation of stereoscopic remote sensing observation for mountains from three perspectives: the three-dimensionality of mountains, stereoscopic remote sensing observation methods, and forward modeling approaches (Figure 1). Our focus on the 3D nature of mountains emphasizes how these features influence observations across the whole stereoscopic space. Stereoscopic remote sensing observation methods encompass various approaches, including observations of vegetation interior 3D structures, tower-based near-surface stereoscopic observations, networked ground observations across terrain gradients, multi-altitude UAV observations, and integrated satellite-UAV-ground stereoscopic observations. Forward modeling underscores both the physical-based parameterization of mountain environments and data-driven, artificial intelligence (AI)-based forward modeling using stereoscopic remote sensing observations.