Discussion - Tailings Dam Performance From USCOLD Incident-Survey Data - Technical Papers, Mining Engineering, March 2001, Vol. 53, No. 3, pp. 49-53- Strachan, C.

Discussion by A.E Agnew Strachan's paper summarizes a wealth of data on water-storage dam "incidents" published by USCOLD and ASCE in 1988, together with a review of tailings-dam incidents published by USCOLD in 1994. The data is summarized by text and 11 histograms. The histograms present a useful compilation of these data, although the patterns within the histogram blocks did not survive the Mining Engineering printing process - so their usefulness is impaired. Unfortunately, the text discussion is too cursory and does not highlight the significant relationships among the various causes of the incidents. Four paragraphs of text and Fig. 1 were devoted to water-storage dams, but the discussion is too sketchy, and Fig. 1 is unclear. Important statistics that could be uncovered are: •the total number of dam incidents tabulated were 500+, •earth embankments constituted 73% of those in operation and •incidents related to earth embankments were 75%, of which dam failures accounted for 24%, major repairs accounted for 42% and accidents accounted for 34%. Strachan concluded that "the majority of the dam failures are associated with overtopping (p. 49), and that the majority of incidents are associated with spillway and other facility structures" - although this is difficult to discern from Fig. 1. Confusion regarding Strachan's use of the terms "dam failures" and "incidents" is compounded by his use of "failures" and "accidents" in one paragraph of his Introduction section followed by his use of "incidents" in the next paragraph. Getting to the meat of Strachan's important paper, the tailings-dam incident survey (USCOLD, 1994), the connection between Figs 2 through 11 seems obscure. First, in the four paragraphs of introductory material, one learns that "incidents consisted of failures, accidents, seepage and other incidents of performance that did not meet design criteria" (p. 50). One does learn the helpful caveat that "structures not related to mill tailings were not included in the review" - so that the Buffalo Creek (West Virginia in 1972) and Aberfan (Wales in 1966) coal-mine failures, together with the ongoing metal-mine remediation work at Summitville (Colorado) do not muddy the waters of his paper. The all-too-brief discussion of the data shown in Figs. 3 through 11 (185 dams in USCOLD plus 26 dams in UNEP) reveals the following: •Figure 3: 57% of the incidents (failures and accidents) occurred with dams less than 20-m (66-ft) in height. •Figure 4: 63% of the incidents were due to failures alone (subset of Fig. 3). •Figure 5: Active dams had 82% of the accidents and 93% of the failures. •Figure 6: Types of dams (upstream, water-retention, downstream, centerline and "unknown") separated according to incidents. •Figure 7: Breaks out the incidents in Fig. 6 by failures, accidents and ground water. •Figure 8: Breaks out incidents by type (overtopping, slope stability, earthquake, foundation, seepage, structural, erosion, mine subsidence and "unknown"). •Figure 9: Separates Fig. 8 data into active and inactive dams. •Figure 10: Separates Fig. 9 data into failures and accidents. •Figure 11: Same as Fig. 8, but broken out by tailings-dam type (Fig. 6). Figures 6 and 7 show that most of the failures are associated with upstream or "unknown" types. Figure 9 shows that the leading causes of incidents for active dams