Professor Reinaldo Garcia from the University of Miami in the USA visited the Institute recently to give a seminar on the topic of Debris Flow. Reinaldo is an old Witonian who stayed at Ashurst from 1997 to 1998, when he was on leave of absence as a Professor at the Central University of Venezuela.

It was during his time in Caracas that Reinaldo started his pioneering research on Debris Flow following the famous and tragic landslide that took place in that region in December 1999. Until then, Reinaldo was renowned in the field of Hydrodynamics and as a developer of several computer models including some simulating the movements of oil slicks.

The 1999 landslide resulted in 15,000 to 25,000 people dead and 20 million cubic metres of alluvial deposits. Many of the landslides occurred on virgin land, ie land with its original vegetation and trees along steep but apparently stable mountain slopes.

The landslide resulted in the destruction of 60,000 houses, some of them built upon old settlements which were also destroyed 50 or more years ago by another flow of alluvial material. The previous disasters however, were not as damaging as the one in 1999 as the number of people affected was much lower.

Reinaldo and his team at the University Central developed a numerical model which allows for the treatment of dry and wet regions. It also includes obstacles in unchanged areas and the associated geometrical irregularities. The fluid behaves like a non-newtonian flow and can be modelled by 2-D equations, including two Newtonian equations, a continuity equation and one for representing erosion. The critical stress was modelled using Egashira’s model. It represents the stress required for the material to start to flow. The programme was based on previously developed shallow water equations software.

The system of equations were obtained using a finite element modelling strategy, the so-called element by element, which does not require the assembly of a huge system of equations. It is a very efficient technique.

The model was tested in the San Julian watershed in the are of interest, ie where the alluvial flow took place. It is a region of steep slopes affected in 1951 by another catastrophic flow. Unfortunately the area continued to develop from then on and became heavily populated just before the December 1999 event. The flow at its peak was 1,000 cubic metres per second. The maximum velocity was around 6 metres per second which results in a high impact factor for the boulders.

The finite element model has more than 3,000 elements, some of them were fixed to represent strong buildings. The initial data was represented by the corresponding hydrograph.

The conclusions were that the flow could be modelled accurately using finite elements. The element by element explicit scheme was found to be economical. Reinaldo applied Egashira’s model and his results could now be compared with those of other models such as the FLO-2D software developed at Colorado State University. The lecture was followed with great interest by the audience and led to a lively discussion.