Detailed rainfall data provided by the Meteorological Office and data for reportable earthwork failures and historical earthwork examination records were also provided by Network Rail. The Data Insights team analysed Met Office data volumes of around 8TB, equivalent to 2.5 million uncompressed photos, to ensure highly robust outputs. The data was processed to build a statistical model to determine the probability that a particular soil cutting would fail during an hour with a given amount of rainfall. The probability that the soil cutting failure resulted in an obstruction on the track was also investigated. 

Frequency of extreme rainfall

The analysis showed that extreme rainfall – defined as rainfall in excess of 10mm in 3 hours – occurred for the equivalent of 35 hours per year for an average soil cutting. As the extreme rainfall intensity increases, the number of hours a soil cutting would experience rainfall of that intensity decreases exponentially.

Extreme rainfall is often clustered both in time and location and is not randomly distributed across the network. 85% of soil cuttings did not experience rainfall in excess of 40mm in 3 hours during the period analysed and, if an earthwork did experience rainfall of this intensity, it was for a sustained period of 4-5 hours on average.

Soil cutting failure rates

The team found that as rainfall intensity increases, there is a disproportionate rise in the number of failures occurring compared with the number of hours when soil cuttings experienced that level of rainfall. 18% of all reportable soil cutting failures occurred during extreme rainfall, but extreme rainfall occurred less than 0.4% of the time.

This failure rate increases gradually with rainfall intensity until a threshold of 45mm in 3 hours is reached, after which the failure rate increases significantly. For example, the failure rate during periods of rainfall over 50mm in 3 hours was greater than 300 times higher than it was for the 10-15mm in 3 hours band.

Failure Risk Categories

The analysis identified the slope height and angle, the catchment area gradient, and the slope face drainage condition as soil cutting parameters that have a significant effect on the probability of failure during extreme rainfall. However, slope face drainage condition was not a significant driver on the probability of failure when rainfall exceeded 45mm in 3 hours.

Different combinations of these parameters have been used to split soil cuttings into eight Failure Risk Category (FRC) bands, to differentiate between high and low risk assets. The soil cuttings in the highest risk FRC band compared with the lowest risk FRC band is 45 times higher for rainfall between 10mm and 45mm in 3 hours, and 10 times higher for rainfall over 45mm in 3 hours.

Probability of encroachment

As rainfall intensity increases, there is an increase in the probability of encroachment, alongside the increase in the probability of failure. The overall percentage of failures that resulted in “obstruction on the line” during extreme rainfall is 63%, over three times the rate for rainfall under 10mm in 3 hours (18%).

Outcome

This analysis has, for the first time, quantified the exposure of individual soil cuttings to different levels of extreme rainfall and the probability of this resulting in (i) a soil cutting failure and (ii) obstruction of the line, based on the size and shape of the asset.

The results of this analysis have been incorporated into RSSB’s R&D project T1269, “Development of a system risk model for extreme rainfall events”. The project aims to develop a Whole System Risk Model  (WSRM) so that the overall safety implications of deciding whether or not to impose a Blanket Speed Restriction during periods of heavy rainfall can be quantified. We are also working with Network Rail to assess whether the Failure Risk Categories that were developed in this analysis, can be used alongside the Earthworks Hazard Category bands to improve the identification of high-risk assets and better target maintenance interventions.