Philip J. Shaller

Estimating the Storm Surge Recurrence Interval for Hurricane Sandy

Hurricane Sandy’s storm surge peaked at 2.87 m (9.40 ft) above mean sea level (MSL) (9.64 ft Manhattan Borough Datum [MBD]) at the southern tip of Manhattan in New York City on October 29, 2012 at 9:24 pm. The peak storm surge coincided with a tide of 0.64 m MSL (0.24 ft MBD), only 30 minutes after a high tide, contributing to catastrophic flooding of near-coastal areas. Traditional flood analysis fitting techniques using various probability distribution functions (normal, lognormal, Gumbel, and log-Pierson III) yielded return frequencies from 667 years to over 10,000 years. The more advanced Lin et al. (2010; 2012) analyses yielded recurrence intervals of between 559 and 650 years for the storm surge alone and 993 years for the surge plus tide. Considering the 2.77-mm/yr (0.11-in/yr) sea-level rise and using the analysis of Lin et al. (2012), future storms equal in magnitude to Hurricane Sandy will result in even greater flooding.

Proposed Revision of Marine Terrace Extent, Geometry, and Rates of Uplift, Pacific Palisades, California

Marine terrace geometries provide a useful means to interpret the rate and style of uplift along emergent shorelines. In southern California, coastal uplift has been linked to activity on blind thrust or reverse faults. Two marine wave-cut platforms were identified in geotechnical investigations at the Getty Villa museum complex in Pacific Palisades, CA. Reconstruction of the late Quaternary geomorphic history at the Getty Villa allows assignment of the upper platform to the ∼320 ka oxygen-isotope Stage 9 Malibu terrace of W. M. Davis and P. W. Birkeland. The lower terrace is interpreted to be the ∼125 ka (Stage 5e) Pacific Palisades platform of J. T. McGill. The wave-cut platforms at the Getty Villa have been uplifted, tilted seaward, and possibly warped. They record late Quaternary uplift at a rate of about 0.3 mm/year and progressive seaward tilting at a rate of about 1°/40 ka since 320 ka. The elevations and geometries of the platforms differ markedly from earlier interpretations, necessitating a reinterpretation of marine terrace geometries throughout much of the Pacific Palisades area. The observed pattern of uplift and tilting suggests the Santa Monica Mountains blind thrust fault of J. F. Dolan and colleagues as the probable source of coastal uplift in this area since 320 ka. Only the central portion of the fault appears significantly active, however. Application of moment magnitude (M) regression equations of Dolan and colleagues indicate that the active portion is capable of generating M 7.0 earthquakes at a recurrence interval of about 6,800 years.

Long-runout landslides and the long-lasting effects of early water activity on Mars: COMMENT

Watkins et al. (2015) propose a clay-lubrication model to explain the extreme runout distances of long-runout landslides in Mars’ Valles Marineris (VM) based on observations of the western margin of Ius Labes, one of the largest VM landslides. Their model proposes that the landslide “overrode and entrained the hydrated-silicate-bearing floor deposits, causing further loss of coherence and permitting the landslide outer zone to spread laterally while moving forward over the low-friction surface.” While this model offers an attractive mechanism to explain the particular outcrops evaluated in their paper, it does not readily conform to the characteristics of long-runout landslides in other planetary settings, the engineering behavior of clays, or the properties of Ius Labes beyond the limits of their study area. The study of long-runout landslides extends back to the late 19 century (Heim, 1882). The unusual characteristics of these landslides have spun off a host of hypotheses, as summarized by Shaller and Shaller (1996). Observations of long-runout landslides in Earth’s oceans (Moore et al., 1989), on the Moon (Guest, 1971), Mars (Lucchitta, 1979), and Saturn’s moon Iapetus (Singer et al., 2012) offer an opportunity to help clarify the issue, assuming these examples are placed into context with the phenomenon at large. In this light, the geologic interpretations advanced by Watkins et al. would have benefitted from a more thorough survey of the relevant literature concerning terrestrial long-runout landslides, as the many morphological similarities between these mass movements on Earth and Mars suggest at least some commonality in mechanism (cf. Shaller, 1991). Among this admittedly voluminous literature, only one study, that of Watson and Wright (1969), cited any field evidence suggestive of a possible role for clay in the long-runout mechanism. This study, which described the Saidmerreh (Iran) landslide, added the proviso that the clay-bearing intervals be saturated to function as an adequate basal lubricant. On Mars, where liquid water is rare, clay lubrication is a priori a more remote possibility than on Earth. Clay lubrication of long-runout landslides is still more problematic for the Moon, which contains no clay (Carrier, et al., 1991). Similarly, Iapetus is too cold to host liquid water (NASA, 2007), which is necessary for clay formation (e.g., Ehlmann, et al., 2011). Moreover, the engineering properties ascribed to the Martian clays by Watkins et al. do not account for the fact that these properties are extremely sensitive to both mineralogy and water content. The authors estimate the VM clay deposits to be 3.6 to 4.0 Ga older than Ius Labes. Clay deposits exposed to the Martian environment over geologic time scales are likely very dry (e.g., NASA, 2013). When dry and compacted, clays are a moderately strong and frictional material (e.g., claystone), with coefficients of friction that range from approximately 0.3 to 0.5 (NAVFAC, 1982). This compares with an apparent coefficient of friction (fahrboschung) of 0.06 for Ius Labes (Shaller, 1991). Coefficients of sliding friction this low are only rarely encountered in terrestrial clay beds, and only when they exhibit optimal mineralogy, a high degree of saturation, and are slowly sheared to residual strength (Watry and Ehlig, 1995). Dry clays are unlikely to have exhibited unusually low frictional properties in the conventional sense and would have been hard to incorporate into the landslide in significant quantities. On Earth, only relatively soft or weak materials are incorporated into long-runout landslides (Shaller, 1991). Finally, the Ius Labes deposit itself provides evidence against a clay lubrication hypothesis. In most places, the light-colored unit interpreted by Watkins et al. to be the basal member of the debris apron is underlain by a lower, darker unit that also appears to be part of the landslide (unit 2 of Watkins et al.), based on the observations that the darker unit (1) has a distinct edge that obscures a neighboring, overridden landslide, and (2) preserves a few longitudinal grooves that are traceable upslope onto the light-colored layer. It is hypothesized here that, instead of representing bulldozed chasma floor material, this exposure actually represents displaced wall rock and that the stratigraphy exposed at the margin of the debris apron is preserved from the headscarp area, a well-known phenomenon in terrestrial long-runout landslides (Shaller, 1991, and references therein that cite 24 published examples). This conclusion is supported by Google Mars imagery of the base of the chasma wall west of Ius Labes, which exhibits the same dark-over-light-over-dark stratigraphy of the landslide’s debris apron. In summary, long-runout landslides are rarely associated with clay deposits on Mars, Earth, or elsewhere. Any clay deposits overrun by Ius Labes were likely dry and would not be expected to exhibit notably low frictional properties. Instead of representing overrun basin deposits, the clay materials exposed along the tapered edge of Ius Labes are interpreted to represent displaced chasma wall material that maintained its relative stratigraphic position in the final deposit. REFERENCES CITED

Rapid in situ conversion of late-stage volcanic materials to halloysite implicated in catastrophic dam failure, Hawaii

Abstract Ka Loko Dam, in Kauai, Hawaii, failed suddenly and catastrophically on March 14, 2006. The resulting breachwas marked by three topographic benches, the lowest of which exposed native volcanic deposits once resident in the dam foundation. These deposits were found to contain outcrops of a waxy, gel-like material that appeared to result from in situ weathering processes. This unusual material was found to be highly enriched in halloysite. Gravel-size pieces in the hydraulic fill of the embankment derived from these materials also exhibited significant in situ weathering and significant halloysite content. Engineers and geologists generally recognize that bedrock materials weather progressively into soil constituents over ‘geological time’, and that this process is accelerated in tropical environments. Still, the strength, stiffness and durability of bedrock, earth and embankment materials are not expected to vary significantly over the geologically short life of a dam. In the case of Ka Loko Dam, however, the volcaniclastic sediments that comprise the local bedrock experienced substantial in situ weathering over its geologically brief 115-year operational lifetime. Prolonged exposure to seepage of anoxic water weathered the sediments completely to saprolite, including weak, sensitive, fine, spherical halloysite.

Analysis of Flood Hazards for a Residential Development

An investigation was carried out to investigate the flood potential at a proposed residential development located in Indio, California. The existing floodplain area adjacent to the project area is designated to receive flood flows originating from Thousand Palms Wash, The Indio Hills and the riverine drainage area along Interstate 10. These flood flows pass through the existing residential development via three flood control channels. A two-dimensional flood routing model (FLO-2D) was applied to study the progression of a flood flow in the interim floodplain area. Model results were analyzed to depict contour plots of the maximum flow depths in the interim floodplain for two discharge conditions that were simulated. The peak discharges entering the project area were estimated from the model results. Two flood control channels are proposed to accept these floodwaters and carry them out of the project area. The U.S. Army Corps of Engineers’ River Analysis System (HECRAS) computer program was used to compute water surface profiles in the two flood control channels for the peak discharges. A sediment transport analysis was also conducted for one of the channels using the U.S. Army Corps of Engineers HEC-6 model. Model results showed that sediment passes through the system well and is not expected to interfere with the function of the flood control channel. Maintenance measures will be in place in order to clean the sediment out of the channel after a large flood event. The design water surface elevations in the flood control channel are based on the greater of either the HEC-RAS calculations using bulked flow, or the HEC-6 computed maximum water surface elevation using the unbulked flow. All sections have adequate freeboard between the design water surface elevation and the top of pad for the design flood event. World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat

The Fire-Flood-Erosion Sequence in California: A Recipe for Disaster

Wildland fires are an inevitable component of many terrestrial ecosystems in California, where alternating episodes of heavy winter rainfall and dry autumnal winds, coupled with extended periods of drought increase the probability of major wildland fires. Factors influencing the intensity and duration of the post-fire hydrologic disturbance include the nature of the local vegetation, burn severity, the geology and topography of the burn area, and the local climate. Many parts of California are underlain by granitic bedrock. Wildfires in granitic terrain result in a common suite of hydrologic aftereffects. The Lowden Ranch Fire took place in an area of granitic bedrock and granite-derived soils in 1999. An investigation of the area in 2002 found only minor evidence of post-fire erosion and sediment transport as a probable result of generally moderate rainfall in the winter of 1999–2000 and the presence of deep, permeable granitic sand in the affected watersheds that reduced runoff. Observations made during our site investigation indicated that by 2002 the area had recovered about 70 percent of its pre-fire erosion resistance. This recovery was due mainly to the reestablishment of ground cover and the breakdown of any hydrophobic layer that had formed in the fire. Comparison of this area with granitic terrains of Southern California burned in 2002 and 2003 and affected by heavy rains in early 2005 indicate that such terrains are relatively resistant to erosion during light to moderate rainfall, but can experience catastrophic erosion during periods of heavy rainfall.

Investigation of Flood and Debris Flow Recurrence: Andreas Canyon, San Jacinto Range, Southern California

An investigation was conducted to evaluate the existing flood and debris flow hazard at the mouth of Andreas Canyon, a major watershed that drains from the rugged eastern slope of the San Jacinto Mountains, California. Unlike archetypical alluvial fans, which form as a result of streams of water spreading sediment and cutting new channels on the fan surface, the Andreas Canyon fan was constructed principally by debris flow processes. Eleven debris flows of varying ages were mapped on the fan surface, ten of which originated in Andreas Canyon. The debris flows are rather large, with a typical volume of about 10 5 m 3 . Nevertheless, hydrologic records available for the watershed suggest that storm events necessary to yield sufficient water to mobilize a debris flow of this size are not extremely uncommon, and should recur every few decades. Archeological records, however, suggest that Andreas Canyon has not experienced a major debris flow in at least 350 years. The absence of debris flows during this period suggests that their occurrence is tied to hydrologic events that are rare or absent in the current climate regime. Details of the type of hydrologic event necessary to produce a debris flow in Andreas Canyon are currently under investigation.

Flood Hazard Analysis and Protection Plan for a Residential Development

An investigation was carried out to develop a flood hazard analysis and flood protection plan for a proposed residential development located in Indio, California. The project area is part of an existing floodplain that straddles portions of the alluvial fans derived from Thousand Palms Canyon and Pushawalla Canyon. Floodwaters emanating from a channel located west of the project area impact the project site. A two-dimensional flood routing model was applied to study the progression of a flood flow in the existing floodplain. Processes simulated included overland flow, infiltration, and bottom boundary roughness. Elevations were based on USGS DEM, rough grading plans, high-resolution LiDAR survey data, and data from a field reconnaissance survey. Model results were analyzed to determine the peak discharges into the proposed development. A channel was proposed to intercept the peak flows crossing the western boundary of the proposed development and convey these floodwaters southward along the western boundary and then eastward along the southern boundary of the project. The U.S. Army Corps of Engineers’ River Analysis System (HEC-RAS) was utilized to predict water surface elevations in the proposed channel, and to establish freeboard conditions with respect to pad elevations. To account for the interception of flows along the western boundary, the flows in the proposed channel were progressively increased to the peak flows. To simulate the outflows across the southern boundary, lateral weir flow was assumed to occur. The results of the analysis show that the proposed flood control channel accept and discharge the floodwaters at the historical locations. The U.S. Army Corps of Engineers’ HEC-6 model was used to conduct sediment transport simulations in the floodplain and onsite channels to determine the potential for sediment deposition. Model results were analyzed to estimate the sediment deposition for each cross section at the peak flow of the event, and at the end of the event. The results of our model studies indicate that sediment is transported through the system well and does not interfere with the function of the proposed stormwater channel.