Ice Stupa Dynamics – Part 1

A good friend of mine has been pioneering an innovative solution to the water scarcity problem of the villages of the high deserts of Ladakh in the Indian Himalayas. The idea is to grow artificial glaciers – ice stupas, named after the Buddhist sanctuaries called stupas – to store the flow-away water during the winter-time for the spring, when water resources are in scarcity for the villagers.


The basic principle is remarkably simple, yet the true physics of glacier-growing represent a multifaceted challenge. I have decided to create a series of descriptive simulation models to illustrate you how will the ice stupas grow and melt. The models are written in the AnyLogic design language, built around JAVA and provided that you have the JAVA plugin installed in your browser, you can try to play with the current model here.


This model demonstrates how temperature will affect the growth and shrinkage of the artificial glacier, which will be essentially a conic ice mountain. Artificial glacier growth is controlled by two parameters, the glacial feed pipe diameter and the glacial flow speed, which is proportional to the altitude difference between the feed pipe head at the real glacier and the location of the artificial glacier.

The glacial feed is coming down from the mountain with a certain debit and gives us the available water to form the artificial glacier. A fraction of it is going to freeze, while the rest will flow away. Of course the freezing fraction will mainly depend on the ambient temperature. The freezing will slowly yield us the artificial glacier, an approximately conic mountain of ice. The growth of the artificial glacier will is dynamic and as we can see it heavily depends on the day-night cycle, with most of the growth occurring during nights, when its colder outside – on average below -10 down to even -30 or less and always below freezing level in the months of January and February. Since we have no melting in these months, this will result in a high net growth of the artificial glacier.

Starting with March, the daytime temperatures will often go above 0, so we can see a melting process starting. Initially the melting only appears during the day, but as we enter into April, melting gets more pronounced and the artificial glacier starts shrinking. Since melting depends on the surface area of the ice cone, melting will be very pronounced in the month of May, when the artificial glacier is still quite big, but the air is already very warm outside. With the current parameter setup, the glacier lasts through the early days of June. In the upcoming videos I would like to refine these to include more sophisticated modellings of the freezing and melting process, as well direct solar radiation. Later on, we will also look into the outflow part to utilize the artificial glacial melt for irrigation.

Please follow the project on Facebook and donate or share the crowdfunding campaign on Indiegogo! It has a been a successful campaign so far with almost a quarter of the project costs raised already, but it needs a big final push for an increased chance of success!

Project description video
Sonam’s Facebook page
Project Facebook page
Project Indiegogo campaign
Model description video
Live model

Made with AnyLogic and JAVA.

Published by Dénes Csala

AI | DATA | ENERGY | SYSTEMS researcher | thinker | modeler | blogger | traveller https://csaladen.es

One thought on “Ice Stupa Dynamics – Part 1

Leave a comment