Volcano mitigation normally means monitoring and evacuation; very rarely does it involve altering a volcano to make it less deadly. There are many and good reasons for not doing any alterations in a volcano, but in some cases it is a good idea.
Gunung Kelud – A brief summary
The relatively small Indonesian volcano of Gunung Kelud (1 731 meters) is known for its powerful and deadly eruptions. It has suffered at least 30 eruptions in the last 1 000 years, 6 of those have been VEI-4 and the eruption of 1586 was a respectable VEI-5. The majority of the other eruptions have ranked in as VEI-3, a size that is quite deadly when you have people living within the risk area.
The main deadly factors for the volcano are lahars (mudflows) and pyroclastic flows. The eruption of 1919 killed more than 5 000 people in lahars, and the eruption of 1966 killed another 200.
Even though the volcano is well monitored and the local authorities are really good at their game, there is a problem with getting the local inhabitants to evacuate before an eruption.
If you have an abundance of large volcanoes and a lot of people living close to them you either become very good at mitigation, or a lot of people will die. The option of moving all those people simply does not exist in Indonesia due to the large population and lack of farmable land. So, first the Indonesians became very good at volcano eruption prediction, and secondly they took hazard mitigation to new unheard of heights. Back in 1919 when the western world did point towards volcanoes saying “Oh, look a Firemountain!” the Indonesians decided to experiment with landscaping their volcanoes to save lives.
After the 1919 eruption it was decided that the most deadly threat was the lahars. And the mudflows from Gunung Kelud were caused by the large crater lake as it overflowed during eruptions. The solution was equally simple as it was ingenious. The Indonesians decided to build a tunnel through the mountain to drastically lower the level of the crater lake.
Not only did it greatly diminish the size of the lahars, it also diminished the abundant pyroclastic flows. As such it was a success, what it did not do was diminish the size of the eruptions.
During the 1966 eruption the tunnel was blocked off by debris, and the eruption also lowered the base of the crater with 60 meters, so a new tunnel at a deeper level was built. In the end it was a feat of engineering as they quarried away in rock and pyroclastic flow debris that was several hundreds of degrees Celsius.
The second tunnel has worked remarkably well. Before an eruption starts the caldera floor bulges and the much reduced lava lake gushes out through the tunnel. This especially worked well during the 2007 to 2008 eruption.
Before the 2007 to 2008 eruption the volume of the lava lake had diminished to below 1 000 000 cubic meters, and as a precursor to the eruption a volcanic dome started to rise out of the lake, as the dome rose the water gushed out of the tunnel stopping any lahars before they could even start. And since there was less water around the interaction between the lava dome and the water was also effectively hindered. In the end the dome cracked open in an effusive manner instead of in the usual explosive way.
It is hard to say if the active mitigation actually changed the eruptive pattern of the volcano, and in a sense of it is pointless to even try to decide that. The tunnels were not built to stop eruptions, or even make them less explosive. The tunnels where built to save lives due to diminishing the lahars. As such the Indonesian endeavor has been a great success. Since the construction only one of the eruptions had a significant death toll (200) and that only occurred after the tunnel had been blocked off by debris. The amount of saved lives is probably into the thousands.
This grand scale experiment with active mitigation should probably save lives if it was implemented in other volcanoes with large crater lakes causing lahars and that have the same problem with large populations close by.