ALPGM (ALpine Parameterized Glacier Model) is a deep learning based regional glacier evolution model designed specifically for simulating and predicting regional glacier evolution. Historical background : This model was developed by Jordi Bolibar with the aim of utilizing deep learning techniques to improve the efficiency and accuracy of regional glacier evolution simulations, particularly in addressing computational efficiency issues when dealing with synchronous simulations of multiple glaciers in a large area. Technical features : Integrating deep learning algorithms to reduce the computational complexity of traditional physical models, using parameterized methods to represent the geometric and dynamic characteristics of glaciers, supporting synchronous modeling of multiple glaciers in a large area, achieving regional scale glacier evolution simulation, combining climate model output data to predict future glacier changes, providing multiple evaluation indicators, and comprehensively analyzing various aspects of glacier changes Core functions : Efficient prediction of changes in glacier area, volume, and mass balance under future climate scenarios, evaluation of spatiotemporal trends in glacier mass balance at the regional scale, quantification of the impact of climate change on mountain glaciers, including glacier retreat rates under different emission scenarios, analysis of changes in glacier water resources and their impact on downstream water resource management, comparison of differences in glacier evolution under different emission scenarios, providing scientific basis for climate policy formulation, identification of glaciers sensitive to climate change in the region, and support for the determination of key protected areas Application case : Prediction of glacier evolution in the European Alps, analysis of glacier water resource changes in the Himalayas, study on the response of glaciers in the Andes to climate change, assessment of global sea level rise contributed by mountain glaciers under different climate scenarios, prediction of glacier changes in glacier tourism areas and management of tourism resources Limitations : Relying on high-quality glacier observation data for model training and validation has limited simulation capabilities for extreme climate events, making it difficult to capture detailed physical processes inside glaciers. The uncertainty of prediction results is affected by the uncertainty of climate model outputs, and simplifying glacier dynamics processes may lead to simulation bias in certain situations Input parameters : Glacier inventory data (geometric parameters such as area, thickness, length, etc.), historical climate data (temperature, precipitation), climate model output data (different emission scenarios), glacier physical parameters (ice density, motion parameters, etc.), terrain data (slope, aspect, etc.) Output result : Time series changes in glacier area, volume, and mass balance, glacier evolution predictions under different emission scenarios, trends in glacier contribution to water resources, spatial distribution of glacier retreat rates, and glacier contribution to sea level rise