Montecito debris flows of 9 January 2018: Physical processes and social implications
Published:May 18, 2020
- PDF LinkChapter PDF
Edward Keller, Chandler Adamaitis, Paul Alessio, Erica Goto, Summer Gray, 2020. "Montecito debris flows of 9 January 2018: Physical processes and social implications", From the Islands to the Mountains: A 2020 View of Geologic Excursions in Southern California, Richard V. Heermance, Joshua J. Schwartz
Download citation file:
Montecito, California, has a complicated Quaternary history of debris flows, the most recent being the Montecito debris flows of 9 January 2018, which were wildfire-debris flow–linked events that took 23 lives and damaged or destroyed several hundred homes. Relative flow chronology, based on boulder weathering, incision rates, and soil dates with limited numerical (radiocarbon and exposure) dating, is used to identify paths of prehistoric debris flows. Topography of debris flow fans on the piedmont is significantly affected by the south-side-up reverse Mission Ridge fault system. Examination of weathering rinds from Pleistocene debris flows confirms that the Rattlesnake Creek–Mission ridge debris flows are folded over the ridge, and that lateral propagation linked to uplift of marine terraces (uplift rate of ~0.5–1 m/k.y.) significantly altered debris flow paths. As communities continue to rebuild and live in these hazard-prone areas, disaster risk reduction measures must take into account both spatial and temporal components of vulnerability.
This field guide includes four stops from Montecito to Santa Barbara. The first stop will be to observe debris flow stratigraphy over the past ~30 ka beneath an earthquake terrace and a prehistoric Chumash site on the beach near the Biltmore Hotel in Montecito. The second stop will be at San Ysidro Creek in San Ysidro Canyon, the site of the largest Montecito debris flow that occurred on 9 January 2018. We will discuss source area and processes of the debris flow, and take a short hike up the canyon to visit the debris flow basin and a ring net designed to reduce the future hazard. The final two stops will explore the debris flow chronology of Santa Barbara over the past ~100 ka. Figure 1 shows the location of the field-trip stops. There is no road log as field sites can be found with a search on a smartphone.
THE MONTECITO DEBRIS FLOW OF 9 JANUARY 2018
The Montecito debris flows of 9 January 2018 were post-wildfire debris flow events that collectively destroyed or damaged several hundred homes and other structures, and claimed 23 lives (Keller et al., 2019). The debris flows occurred several weeks after the Thomas Fire had burned the hills above Montecito, California.
The flows came as a complete surprise to many people of Montecito. Most had never heard the term ‘debris flow’ and had little idea of what to expect. Many residents of Montecito had large boulders in their gardens and liked them, but they had little or no idea where they came from. The volume of debris flow boulders and mud is not well measured. Recent estimates from the U.S. Geological Survey (Kean et al., 2019), based on average depth of the flows, are ~765,000 m3 (~1,000,000 yd3) from the canyons above Montecito and another 153,000 m3 (~200,000 yd3) from Santa Monica Canyon above Carpinteria. The majority of the total volume in Montecito came from three canyons (Montecito, San Ysidro, and Romero). The University of California Santa Barbara debris flow research team is currently measuring the amount of debris deposited from the 2018 debris flow, using before and after LiDAR measurements (Keller et al., 2019).
Debris basins exist in most of the larger canyons above Carpinteria, Montecito, Santa Barbara, and Goleta. The purpose of the basins is to trap sediment from debris flows and floods. Several ring nets consisting of large interlocked steel rings have also been installed across some of the canyons to trap debris moving down the canyon. The size of the debris basins varies from a few thousand m3 to a few tens of thousands of m3. The ring nets add a few more thousand m3. The debris basin in Santa Monica Creek is the largest, ~159,000 m3 (~208,000 yd3) (Santa Barbara County Flood Control and Water Conservation District, 2017). The debris basins are, with the exception of Santa Monica Creek, too small to trap the large debris flows that occurred in 2018. The debris basins were not destroyed, but, in 2018, the debris basin dams across the canyons partially blocked the flows. The debris basins filled quickly, and the dams (which can become obstructions to flows) caused the debris flows to flow over the top of some of the dams by several meters. The flows forced over and around the debris flow dams spread laterally below the basins, contributing to downstream inundation of mud and boulders. The source of the mud was the hillslopes, and the source of the boulders was the canyon floor and channel bed.
The magnitude (M) of a debris flow is the log-base 10 of the volume of the flow in m3 (Keaton et al., 1988). The ART (average return period, time between events) of a large debris flow (M5+) from any one canyon, based on conditional probability, is 500–1000 yr (Keller et al., 2019). Assuming the annual probability of a wildfire at a particular basin following recovery from a previous fire rated at 0.10 (1 in 10), and that the probability of a short duration high-intensity storm is 0.02 (1 in 50), then the conditional probability is 0.002 (1 in 500). If the probability of a high-intensity storm (such as Montecito 9 January 2018) is 0.01, then the conditional probability is 0.001(1 in 1000). This would suggest that high magnitude debris flows emerging from any one canyon are relatively rare (Keller et al., 2019). However, because each fire and flow is an independent event, the time may be shorter or longer between any two events. Smaller events (debris flows and debris floods) have an ART, perhaps as short as 20–30 yr, which is the ~ART of wildfire (Kean et al., 2019). Small debris flows often occur in the first several years after a fire, and several in recent decades have caused damages to property. However, the 2018 Montecito events were much larger than past historical debris flows. A debris flow in 2019 occurred in Romero Canyon above eastern Montecito (Keller et al., 2019). That event, while significant, was much smaller than the 2018 event. The debris flow basin with a capacity of ~21,000 m3 (~27,000 yd3) functioned well by capturing boulders and mud, and reducing potential down canyon and piedmont damages.
Paths of debris flows in Montecito are maintained across the relatively low topographic expression folds and young fault scarps of the Mission Ridge fault system (Fig. 2). Where fault scarps are present, flows narrow and flow through incised channels (water gaps).
Debris flow fans are prone to channel avulsions caused by radial fan channels being blocked by boulder deposition and changing channel morphology. Thus, as the fan evolves over time, debris flows often take different paths.
Perception of the debris flow hazard in Montecito and Santa Barbara has forever changed as people have become aware of the debris flow hazard. The boulders are now an iconic reminder of the 2018 event. Large boulders are also common within the city limits of Santa Barbara. A goal of our research is to better educate people about the potential hazard; Santa Barbara has at least the same hazard risk as Montecito, as do parts of Goleta, just west of Santa Barbara (Fig. 2).
WILDFIRE-DEBRIS FLOW CYCLE
The linkage between fire and debris flows may be depicted as a periodic cycle we call the wildfire-debris flow cycle (Fig. 3). The cycle is discussed in more detail in Keller et al. (2019), and what follows is summarized from that source. The cycle is a process-response model, where the response may be a fluvial flushing of sediment in response to low to moderate intensity of rain, or a large debris flow or series of smaller flows in response to rainfall of high intensity, or a large debris flow followed by flushing events following moderate precipitation events. The size of a debris flow depends on the rainfall intensity and duration, size of the basin, availability of sediment (boulders) that could contribute to another debris flow, and how the hillslope hydrology and vegetation have changed over time.
The first part of the wildfire-debris flow cycle is a wildfire, which, in southern California, recurs on time scales as short as 20–30 yr (Mensing et al., 1999). With climate change and human interference with fire processes, the fire intensity is apparently increasing, and the fire season is now nearly the entire year (IPCC, 2014; Mann et al., 2016). Wildfire plays an important role in that it removes vegetation and changes the composition and texture of surficial materials, resulting in an increase in both runoff of water and sediment.
Precipitation is the second process of the wildfire-debris flow cycle. The occurrence of a high-intensity rainfall within the first months to a year or so following the wildfire, prior to recovery of the vegetation, increases the likelihood of a debris flow. Debris flows in southern California tend to occur in response to short duration, intense precipitation events of less than 30 min. The best correlation between intensity of rainfall and debris flow is with a 5 min rainfall intensity threshold, producing a debris flow that follows the peak in precipitation by only ~5 min (Kean et al., 2011). The 2018 Montecito debris flows began several minutes after an intense burst of rain of ~15 mm over ~5 min, which moved quickly west to east across the Montecito foothills and piedmont.
Hillslope erosion is the third process of the wildfire-debris flow cycle. Most debris flows in the chaparral of southern California are linked to surface erosion. Runoff and erosion increase following wildfire because soils are more impervious due to: development of a waxy layer near the surface that retards infiltration of water; infiltration of ash and fine particles into the soil that seals the surface; and lack of protective cover for soil from intense precipitation (Wells, 1981, 1985).
Erosion of rills on slopes in the source area of debris flows (often long, narrow, closely spaced surface channels eroded into slopes on burned areas) has been the subject of intensive research (Schumm et al., 1984; Nearing et al., 1997; Favis-Mortlock et al., 2000; Yao et al., 2008). An early study that related the formation of rills to runoff and debris flows is Wells (1981). Wells was in the mountains above Los Angeles, performing experiments on slopes by artificially raining on slopes that he had burned. He noticed that with intense precipitation, hillslope rills developed quickly, and each rill had a small debris flow, some of which coalesced into larger ones.
Rills occurred over large areas in the source area for the 2018 Montecito debris flow. The density of rills and the runoff and production of fine sediment contributed to the generation of mud. The intense precipitation preceded the debris flow by only a matter of minutes. Details of hydrologic processes linked to sediment transport, rilling, and small debris flows are being studied following the 2018 Montecito debris flows, and it is apparent that generation of the mud is a complex process.
The fourth process of the wildfire-debris flow cycle is a debris flow and/or a fluvial flushing event. The process for debris flow involves relatively fine sediment (ash, silt, sand, and gravel) eroded from slopes into channels where boulders are in storage. Mobilization of the sediment into a fast-moving debris flow is conceptually understood. However, details of the transformation are poorly understood (Kean et al., 2013). Nearly every chaparral wildfire with subsequent precipitation that forms rills produces small debris flows of M1–3. The occurrence of high magnitude M5+ events are infrequent, and details of the processes that transform boulders several meters in diameter stored in the channel, to a moving mass of boulders traveling at high speed that are deposited as boulder fields and levees are poorly understood.
Fluvial flushing may be part of the fourth stage of the wildfire-debris flow cycle. Sediment flushing may occur with or without high magnitude debris flows or during moderate precipitation after a high magnitude debris flow has occurred (Florsheim et al., 1991, 2017; Keller et al., 1997 and 2019). Flushing followed the 2018 Montecito debris flow. The flushing was produced by moderate rainfall intensities of ~15 mm/hr.
The wildfire-debris flow cycle emphasizes that wild fire and debris flows are closely linked by intense precipitation. However, occurrence of a high magnitude debris flow is also a function of geomorphic instability. The instability is the result of the accumulation of boulders and other sediment in source area channels following a debris flow. The time necessary to build up sufficient material to produce another large debris flow is largely unknown, but probably is hundreds to thousands of years. If a wildfire is followed by an intense storm that occurs when geomorphic instability is high (that is, there are many boulders stored in source area channels), the possibility of a large debris flow is increased. Lacking geomorphic stability, sediment flushing or flash flooding is more likely (Keller et al., 2019).
DEBRIS FLOW INITIATION AND RILLING
During high-intensity rainfall events, concentrated flow from runoff leads to the generation of rills. Following wildfire, extensive rill networks that are lined with debris levees have been reported to occur on hillslopes on gradients >0.4 and are often associated with runoff generated debris flows (Wells, 1981; Parrett, 1987; Cannon et al., 2001; Nyman et al., 2011). Levee-lined rills have very high sediment concentrations, occur near the tops of ridgelines, and progressively widen and deepen downslope (Wells, 1981; Parrett, 1987, Wohl and Pearthree, 1991; Meyer and Wells, 1997; Cannon et al., 2001). Earlier researchers have described and noted their importance. Wells (1981) was the first to hypothesize that rills were a primary contributing factor in generating debris flows, and he described rills lined with levees as numerous tiny debris flows that occur on hillslopes in a matter of minutes. Meyer and Wells (1997) describe shallow levee-lined rills (~1 cm) that gradually incised to a depth of up to 12 cm downslope, and concluded that the increasing sediment concentration from levee-lined rills resulted in the transition to debris flow. They have since been observed around the world (United States, Australia, France, and Italy) on a wide-variety of lithologies (shale, sandstone, limestone, granite, schist, gneiss, and unconsolidated fanglomerates) (Wells, 1981; Parrett, 1987; Cannon et al., 2001; Gabet, 2003; Langhans et al., 2017). Blijenberg et al. (1996) and Langhans et al. (2017) have reproduced them in the field and describe their formation. Recently, Langhans et al. (2017) classified Wells’ (1981) levee-lined rills as a distinct process he called “hillslope debris flows,” and provided a detailed description of their morphology and initiation, as well as a conceptual model for gravity-runoff-driven granular flows. The distinguishing characteristic between levee-lined rills and their water-rich counterpart is the nature of their flow. “Rills” are water rich, have lower sediment concentrations, and are governed by fluid flow, whereas “levee-lined rills” recruit enough sediment to transform the dilute suspension into one which has viscous behavior and is governed by granular flow. The specific mechanism by which this transformation occurs varies between interpretations and between regions, and, though they are distinctly different from water-rich rills, they are capable of generating debris flows. However, while their morphology is well documented, there exists a lack of measurements of levee-lined rill geometries and volumetric contributions at the hillslope and catchment scale. In addition, there is minimal theory available to describe their formation, network patterns, and how incision and mixing of cohesive subsoil occur.
Following the 9 January 2018 debris flows, extensive networks of levee-lined rills were observed in the upper parts of the watersheds, and were hypothesized to be the primary mechanism by which the flows were generated (this study). Since then, our research objectives have been to (1) improve our understanding of levee-lined rill network formation, (2) measure the volume of mud generated from rilling, and (3) understand how levee-lined rill network formation and volumes vary with topography and lithology. These objectives are being addressed using Structure from Motion (SfM) orthomosaics and digital elevation models (DEMs) generated from high-resolution aerial photographs that were taken concurrently with the airborne LiDAR (flown ~2 wk after the event). Orthomosaics and DEMs with 5 cm resolution were used to identify processes and spatial patterns of eroded material, and to make systematic measurements of rill geometries and patterns, in order to estimate volumetric erosion.
SOURCE AREA OF THE 2018 MONTECITO DEBRIS FLOWS
The source areas for the debris flows are visible from most of Santa Barbara, and we will view them as we stop at various places on the debris flow fans and the beach. Rilling occurred on both the alternating and southward-dipping sandstone and shale formations (expressed as zones of varying topography within the mountain range), but mapping shows that their abundance is clearly controlled by lithology (Fig. 4). Rill networks were abundant on shale formations, and, despite similar rainfall and wildfire conditions, rill occurrence was relatively scarce on sandstones, where the most extensive ground disturbances were gullies and shallow colluvial failures. The absence of levee-lined rills on sandstone units reflects the lack of soil cover and significant soil armor from large angular blocks of colluvium. Continuous armoring prevented rilling in some places, and on these steeper, more rugged hillslopes, rilling was replaced by shallow (<5 cm) anastomosing flow concentrations without well-defined banks and sheetwash (Fig. 5). These flows were interrupted and dispersed by boulders and bedrock protuberances.
On hillslopes with continuous soil cover, rills were sub-parallel, evenly spaced (~1 m apart), cut across contour on the burned hillslopes, and exhibit average confluence and bifurcation angles of 34 degrees. Rill and levee cross-sectional geometry, as well as rill spacing, both increased as power functions of contributing drainage area per unit width and local slope (Fig. 6). Rill density decreased downslope due to merging of secondary rills and widening of primary rills (Fig. 7). Mean volume estimates on 100-m-long slopes for an individual rill were similar between shale and unarmored sandstone slopes (~4 m3), and were consistently larger on concave planform hillslope geometries. Mean levee volumes were 11% of rill volumes for both shale and sandstone units. Volumetric estimates of rill erosion were 363,300 and 257,000 m3 for all five watersheds that produced debris flows, corresponding to ~30%–40% of fan volumes measured by Kean et al. (2019).
Levee-lined rills in the Juncal Shale (the uppermost shale unit) contributed a substantial amount of well-mixed sediment and water to 0 and 1st order channels, which resulted in boulder recruitment from the sandstone units below. The stream channels were subsequently scoured (seen at San Ysidro debris basin), which increased the volume of the flows substantially. This result emphasizes that lithology and topographic position are key variables in determining the potential for rilling of soil-mantled hillslopes to generate mud for post-wildfire debris flows. Information regarding rill patterns and geometries as they progress downslope on various lithologies, and their contribution to the total sediment budget can serve to improve quantitative models and parameterizations of rilling within models that aim to forecast debris flow potential and volume.
We define boulder geomorphology as the study of boulder morphology, weathering, and chronology. A weathering rind (often a red-brown layer) on old boulders is the result of mineral weathering and mechanical fracturing of the outer few millimeters of the boulder. Over thousands of years, these rinds are transformed to case hardening at the surface, which eventually may be several cm thick (Dorn et al., 2017).
Boulders on the surface of a debris flow (as a landform, such as a debris flow fan) are concentrated in boulder fields or boulder levees, but scattered boulders may be found anywhere on a flow surface, even near the distal parts of a flow. Scattered boulders deposited on the surface of the 2018 Montecito debris flows are present on lower parts of the flow near the ocean, where the dominant deposits are from the wet-tail of the flow (mostly mud flow). The 2018 boulders are generally angular and lack signs of weathering. It is assumed that the transport and deposition during a debris flow removes most of the evidence of past weathering. Boulder geomorphology (Keller et al., 2019) includes the relationship between thickness of weathering rinds and time to delineate a calibrated chronology (Figs. 8–12).
Most of the piedmont of Montecito, Santa Barbara, and Goleta is composed of debris flow deposits that form debris flow fans. Based on boulder geomorphology, individual large debris flow paths can be mapped from the source area to the sea. Boulder morphology chronology, along with active tectonic processes, is also helping us understand the history of the flows with implications for defining exposure of the hazard over time and space.
This field guide includes several stops to observe debris flow deposits in Montecito and Santa Barbara that range in age from ~100,000 ka to the 2018 event. During lunch, we will discuss our social science research that examines some of the human dimensions of the debris flows. Ages of debris flows are based on weathering rind/case hardening thickness linked to C14 analysis, exposure dates, incision rates, and soil dates. This part of our research is in an early stage of development. The hypothetical chronology is shown in Figure 12. The weathering rind thickness appears to fall into three classes: thin, intermediate, and thick (see Figs. 9–11). These thicknesses may correspond to climate cycles (Oxygen Isotope Stages [OIS] 1, 3, and 5) when sea level was relatively high, temperature was warm, and intense precipitation and chaparral vegetation that could support a strong wildfire-debris flow cycle were more likely (Mensing, 2014).
Stop 1: Between the Mouth of Montecito Creek and the Biltmore Hotel at 1260 Channel Dr. (Olive Mill Rd. exit on U.S. 101) (best visited at low tide)
This site (34.4173° N, 119.6420° W) is located east of the Biltmore Hotel beyond the Coral Casino Beach Club and adjacent condominiums, just east of Hammond’s Beach. Just east of the site Montecito Creek enters the ocean.
During some past floods and debris flows and during the 2018 debris flow, the present main mouth of Montecito Creek was not the only distributary channel of Montecito Creek. The 2018 debris flows mostly bypassed the creek channel and debris basin on the lower piedmont and flowed down Olive Mill Rd. Some of the flows inundated the Montecito lower village, damaging hotels and other structures. Flows continued across Highway 101, and the freeway was covered in more than 3 m of mud and debris. Flows continued across Highway 101 and down Channel Dr. between the Biltmore Hotel and the Coral Casino Beach Club, filling a below-ground parking structure with mud. Several homes were damaged (one destroyed) south of Highway 101.
It was reported that Mission Creek took a new course during the 2018 debris flow, but did it? Figure 13 is a map of the area from 1869. There were two channels that flowed to the ocean in 1869 (U.S. Coastal Survey, 1869). This is typical behavior of a debris flow fan where channel avulsion is common. The eastern channel (main channel of Montecito Creek) received little flow in 2018 by comparison to the western channel, which is along Channel Drive adjacent to the Biltmore Hotel. The western channel no longer has the expression of an active channel, as it was filled and paved over when developed. However, during recent floods the western channel has delivered flood flow to the ocean.
The western channel is where the 2018 debris flow entered the ocean. An older flow of ~1 ka may also have entered the ocean at the western channel. Boulders, nearly on the beach in front of the condominiums, have weathering rinds similar to those at Rocky Nook Park in Santa Barbara, dated at ~1 ka. A low sea-cliff exposure is located below a 2–3-m-high wave-cut platform dated at ~2 ka (14C on shell carbonate). Sea level has been nearly constant the past several thousand years, and, therefore, the platform was likely produced in one uplift event, forming an earthquake terrace. The earthquake was likely ~M7 to have produced 2–3 m of uplift (Wells and Coppersmith, 1994).
The stratigraphy of the site is shown in Figure 14; a prehistoric Chumash site at Hammond’s Meadow is exposed at the surface. Two older debris flow events (dated at 16,728 ± 252 yr B.P. and 28,943 ± 389 yr B.P., respectively, are present. The younger event is likely a mud flow, and the older is a boulder-rich debris flow. On top of the younger flow and below the wave-cut platform is another gravel deposit that may be a sediment flushing event. Thus, this site contains perhaps four debris flows and one sediment flushing event.
Stop 2: San Ysidro Creek (Debris basin is located ~400 m upstream from the San Ysidro Trailhead at E. Mountain Dr.)
At this stop (34.4498 ° N, 119.6227° W), we will observe several aspects of the large flow that came down San Ysidro Creek, causing damage to homes and loss of lives on 9 January 2018. The flow had an estimated velocity of ~9 m/s. Flow marks on trees are present, and original parts of the boulder field remain (Tom Dunne, 2018, personal commun.). The boulders have no weathering rinds, and we assume that any rinds present on boulders in the stream bed were removed by the boulders moving in the flow.
The San Ysidro Creek debris basin (capacity of ~8000 m3) overflowed in 2018 and caused the flow to widen and contribute to damage downstream. The San Ysidro debris basin, as with others on the larger canyons above Montecito, is periodically cleaned out. However, the basins have small capacities compared to large debris flows (M5+). We will take a short walk up the canyon beyond the debris basin to observe a ring net installed across the canyon floor to trap debris, as well as to observe debris flow features such as canyon scour and deposition.
Stop 3: Mission Anticline (American Riviera) at Orpet Park (34.4379° N, 119.7048° W): Alameda Padre Sierra
Orpet Park is a small park located ~1 km east on Alameda Padre Sierra from Mission Canyon Rd. where E. Los Olivos St. (the road by the Old Santa Barbara Mission) changes its name to Mission Canyon Rd. The park is on the south flank of the Mission Ridge anticline. Just north of Alameda Padre Sierra S, there is a debris flow boulder field. Boulders of similar ages (based on boulder morphology) are located on both the crest and north flank of the anticline (Figs. 9 and 15—sites 20–24). Given that OIS Stage 5 lasted ~40,000 yr, several debris flows must have occurred. These flows formed one of the largest debris flow fans in Santa Barbara, referred to as the Rattlesnake Canyon debris flow fan. Thus, the Mission Ridge anticline formed in the past ~100 ka. The debris fan apex is in the Santa Ynez Mountains above Santa Barbara, and extends through what is now the city of Santa Barbara. That fan was then folded to form Mission ridge. Boulders of similar age (~100 ka) are found south of the ridge in the city of Santa Barbara. Most boulders have been cut into blocks used in the construction of the many rock walls in the area. Thus, when we see abundant rock walls, we are often in old debris flow sites. The folded debris flow fan deposits provide a good estimate of the age of the uplift and folding of the ~100 m of relief over the ridge (Gurrola, 2005; Melosh and Keller, 2013; Gurrola et al., 2014).
Stop 4: Rocky Nook Debris Flow (34.4409° N, 119.7126° W), Mission Park, and the Mission Rose Garden and Historic Buildings
The headwaters of Mission Creek and its major tributary, Rattlesnake Creek, are high in the Santa Ynez Mountains above the city of Santa Barbara. Rattlesnake Creek joins Mission Creek just upstream of the Santa Barbara Museum of Natural History, and the waters of Mission Creek have, for at least several hundred thousand years, delivered sediment to what is now the urban area of the city of Santa Barbara. The city, along with much of the piedmont of Carpinteria to west of Goleta, is built upon debris flow fans, lobe-shaped deposits consisting of stream gravels and debris flow deposits. With respect to Mission Creek, erosion is concentrated in the Santa Ynez Mountains, and the eroded sediment, from silt to boulders, is deposited on the fan. A few tens of thousands of years ago, Mission Creek flowed nearly north-south to the Pacific Ocean through what is now the city of Santa Barbara. Abandoned channels that we refer to as paleochannels of Mission Creek are common in the city (but are now filled in and mostly paved over). The most prominent of these flowed into El Estero, which was a large lagoon extending from approximately the location of Santa Barbara High School to the ocean. Long before the Chumash people and the Spanish arrived, the stream was diverted by geologic processes (uplift and folding) from near the Natural History Museum to its present course, ~700 m west of the museum. From there, the stream flows south through Oak Park and then to the southeast to the Pacific Ocean. El Estero, or the lagoon, denied its nourishing waters from the creek, developed into a salt marsh, only occasionally receiving floodwaters from Mission Creek. The salt marsh was partly filled with debris from the 1925 Santa Barbara earthquake and paved over. Why was Mission Creek diverted to the west from its more southerly route to the ocean? The reason for this seems to be the development of Mission ridge, which blocked the path of Mission Creek. Mission ridge is a geologically young anticline that apparently is growing westward as a result of uplift and associated earthquakes. This uplift blocked Mission Creek, diverting it to the east several tens of thousands of years ago (Keller et al., 1999; Melosh and Keller, 2013).
Mission Creek used to flow across the west end of the anticline near the Old Mission of Santa Barbara; however, uplift defeated the channel, causing the channel to subsequently flow along the base of the westward extension of the north side of Mission Ridge near the Santa Barbara Historic Museum. The linear hill or ridge is a “fold scarp” produced by displacement on the buried south-dipping Mission ridge fault). Mission Creek cannot flow directly southward (as it once did), because Mission ridge blocks the stream from flowing in that direction. As a result, Mission Creek must flow to the west around the end or “nose” of the fold. The fold scarp that Mission Creek follows along the northern base of the hill is remarkably linear, which suggests that geologic structure and process are playing a significant role in stream channel location. The next part of our journey will take us to Rocky Nook Park, where we will discuss in more detail the origin of the large number of boulders found there.
Rocky Nook Park
The numerous large boulders scattered about on the surface exemplify the uniqueness of Rocky Nook Park. These boulders (as with those at the museum site to the immediate west) are composed of sandstone from the Santa Ynez Mountains. The boulders represent an ~M6 debris flow that occurred sometime in the recent geologic past. The volume of the flow deposits is ~5–8 million m3 or 1–2 million dump truck loads. This is several times the volume of the 2018 Montecito debris flows.
At Rocky Nook Park, there are several debris flow features, such as flow lobes and levees (Selting, 2001). Figure 16 shows the debris flow through Rocky Nook Park upstream to the landslide that apparently is the source of the flow. Some of the boulders have an open framework structure consisting of boulders piled on top of each other, where void spaces between the boulders have no fine sediment between them. The boulders are transported by the flow, in part because the unit weight of the flowing mud and the boulders is nearly the same. The boulders tend to move near the top of the flow, pushed to the front and sides. The flow generates dispersive stresses proportional to the square of the shear stress that causes large boulders to migrate toward the top and sides of the flow, where dispersive stresses are lower (Bagnold, 1954). Often, the largest boulders in a debris flow are on the surface or on the margins of the flow, where they may form linear piles of open framework boulders that are termed “debris flow levees.” Finer particles deposited initially between the boulders are not present. Debris flow levees are found in Rocky Nook Park and several miles up Rattlesnake Creek. The boulders are geologically young. Weathering rind thicknesses of these boulders (see Figures 9–11, site 3) are ~4 mm, compared to no rinds present for the 2018 Montecito flows. The debris flow that produced the boulders probably occurred a year or so following a wildfire in response to a high-magnitude, intense rainstorm (see wildfire-debris flow cycle; Fig. 3). Large landslides occurred upstream in Rattlesnake Canyon, probably forming temporary natural dams across the creek. As these dams were overtopped and eroded, mud and boulders “burst out” and carried the boulders downstream as a debris flow. The boulders are deposited when the flow spreads out and disperses in the down valley direction. In other words, we hypothesize that the debris flow probably originated in part from the failure of a landslide dam that blocked Rattlesnake Creek. A young landslide of about the right size is present at Skofield Park (Fig. 16), recognized by the steep horseshoe shape of the east canyon wall, and the relatively flat blocks of land that form playing fields (Selting, 2001; Urban, 2004; Gurrola et al., 2016).
The debris flow that came from Skofield Park ~1000 yr ago (Urban, 2004) went through Rocky Nook Park and the Santa Barbara Museum of Natural History, and then turned south to the ocean. If such an event were to occur again today, many homes and buildings, including the museum, would be destroyed, and the loss of life would be catastrophic.
While in Rocky Nook Park, you can observe the hill across the stream to the south. If you look carefully, you may observe the south “limb” of the gravels (debris flow deposits) that form part of the Mission ridge anticline (Gurrola, 2005).
Leaving Rocky Nook Park and turning to the left (south) and walking across the Mission Creek bridge, we are walking along a wind gap of Mission ridge. At the bridge, the creek is trending nearly east-west, due to the diversion caused by Mission ridge, which has been uplifted and folded by a series of earthquakes during the past ~100 ka. Keeping on the east side of the water gap through Mission Canyon, and carefully crossing Mountain Drive and Alameda Padre Sierra, we enter Mission Park.
Mission Park (34.4409° N, 119.7126° W)
Mission Park, located just east of Mission Santa Barbara, is an interesting landscape in that the park in its entirety is a prehistoric (paleo) channel of Mission Creek that flowed south into prehistoric El Estero (lagoon) in lower Santa Barbara. As you walk through the park, it’s easy to imagine that this was a stream channel that flowed south, and then turned west through what is now the rose garden of the mission, before turning south again and eventually into the Santa Barbara High School area. As you walk along the channel at the park, imagine what it might have been like with water flowing in the channel. We observe in several locations the tops of boulders poking above the bottom of the channel, suggesting that a boulder bed stream is below. Other boulders on the eastern channel margin, and incorporated into the jail house of the old mission, and just east of the rose garden, have weathering rinds thicker than the Rocky Nook flow (Figs. 9–11, site 18). Notice, however, that the landscape in Mission Park is quite different from Rocky Nook Park. Nowhere do we see the large boulders piled around on the landscape that are so obvious at Rocky Nook. We conclude that the large debris flow at ~1 ka that came down Mission Creek did not go down this paleochannel of Mission Creek—otherwise, we would see similar boulders here today. Thus, the diversion of Mission Creek to the west occurred before the Rocky Nook debris flow. The youngest debris that did move through the present wind gap is at least several thousand years older than the Rocky Nook flow.
The Rocky Nook debris flow does not end at the park. The flow moved west as it was blocked by the growing Mission ridge for nearly 700 m before Mission Creek, and the Rocky Nook flow turns south through another gap in the Mission Creek fault system (Keller et al., 1999; Keller and DeVecchio, 2013). The Rocky Nook debris flow can be traced by boulder geomorphology south to near U.S. 101. In the future, when another large M5+ debris flow occurs in Mission and/or Rattlesnake Creeks, the flow will inundate the western part of the city of Santa Barbara to the ocean. It is not a matter of if, but when.
Lunch Discussion: A Disaster of Social Vulnerability
The likely lunch stop is Manning Park, Montecito, which is on an actively growing anticline with a water gap (Oak Creek), where a moderate ~M4 debris flow occurred in 2018.
As more people live in hazard-prone areas, debris flows increasingly enter into existing sociocultural contexts, urban environments, and relationships of privilege and disempowerment (Santi et al., 2011). The debris flow that took place in Montecito on 9 January 2018 is one such example, and can be considered a disaster resulting from social vulnerability” (Hewitt, 1997). Roughly one third of the fatalities consisted of working-class immigrant families in a community where only 9% of the residents are non-white. The majority of Montecito residents are wealthy (66%), making more than $150,000/yr, and most residents (63%) hire at least one worker at their property, including gardeners, housekeepers, caretakers, babysitters, and home care for the elderly.
Assessing vulnerability is a methodologically complicated task (Birkmann, 2013), and the tools available are not adequate for understanding the unique case of Montecito. This is partly because broad quantitative vulnerability assessments consider the spatial extent of the community, but do not consider the local socioeconomic heterogeneity. General assessments have no mechanism for including the marginalized populations within a community, and these are usually the most vulnerable (Wisner, 1998)—for instance, the homeless during the Tokyo earthquake (Wisner, 1998) or informal workers during the Thomas Fire in Ventura and Santa Barbara Counties. Broad quantitative vulnerability assessments also fail to capture the temporal dimension of vulnerability—that is, how vulnerability can change over time as collective memory fades.
In order to understand community vulnerability in Montecito, an interdisciplinary team of researchers at the University of California, Santa Barbara, employed a combination of quantitative and qualitative social science. Within a few months, Montecito residents were mailed a survey designed by quantitative social scientists. Within the same timeframe, a team of qualitative social scientists conducted 22 in-depth interviews with residents, workers, grassroots’ organizations, and public authorities. The interviews allowed for maximum flexibility, and interviewees were given space to communicate their own stories and experiences, ushering in new topics that might have otherwise been overlooked. These efforts enabled researchers to consider a wider range of the community, its local dynamics, and the particularities of geography that shape both spatial and temporal components of vulnerability. The sections below illustrate themes or dimensions that would have been overlooked by a broad quantitative vulnerability assessment.
The Hidden Renter Class and Invisible Labor Force of Montecito
When interviewees were asked to reflect on perceptions of Montecito as a wealthy community, the responses were largely critical of such a notion, arguing instead that the community is “a hodgepodge” consisting of working-class families and “a huge percentage” of renters who relocated to offer their children a better public education. One respondent described herself and family as impoverished, “surrounded by people that are not in the same class as the other.” Another remarked, “Montecito isn’t what most people imagine, this enclave for billionaires, you know? It’s a lot of ordinary people.”
Historical Amnesia and the Temporal Dimension of Vulnerability
One of the key factors that rendered Montecito residents vulnerable was a lack of knowledge about debris flows and previous events in the region. As one long-term resident noted: “Anyone who’s lived in Montecito for more than ten years knows where the water goes.” Yet many residents were new to the area or had no experience of wildfire followed by intense rainfall, including one resident who recently moved from Los Angeles and stated: “We didn’t expect this to happen. Even when they said the rains were coming, you don’t expect a mountain to come down.” In the aftermath of the disaster, most residents were understandably in a state of shock, having never seen or heard of a debris flow. Some felt more familiar with terms like “mudslide,” “landslide,” and “flash flooding,” and some even felt that these events were particular to places in the Global South (low and middle income countries located in Asia, Africa, Latin America, and the Caribbean). As one resident noted: “I know flash flooding and landslides happen … in foreign countries … but the mud and debris flow? This is new terminology for me.”
Failures in Communication
The internal dynamics of communication and governance proved to be an important factor that contributed to vulnerability. In the aftermath of the debris flow, baffled by the lack of warning, many questioned the ill-drawn evacuation lines enforced by the county. Evacuation lines were east-west (movement of fire) for a mostly north-south hazard (the debris flow moved downstream along stream corridors). This mirrored the mandatory and voluntary evacuation zones established for the previous evacuation due to the Thomas Fire. Yet, as many were quick to realize—some unfortunately too late—water flows downhill. It doesn’t stop at an imaginary line. In the few days between the Thomas Fire and the debris flow, one survivor we interviewed recalled:
I looked up at the mountains with [my husband] … bare with rock … and I said, “Oh my god, where are those going to go?” I still have chills remembering that. This is right after the fire, and just having this kind of premonition. He said, “No, don’t worry about it. They’re fine, they’re big huge boulders.” I said, “Exactly, they’re big huge boulders and what’s holding them in place now after this?” And the mountain was just scarred black and red and bare. And it just looked like a different landscape.
Public authorities did not adequately educate the community about past debris flows and what they are, and they made a poor choice by choosing the same evacuation line as used during the Thomas Fire. On the other hand, authorities have been very supportive of residents in the after phase, creating the Montecito Center for Preparedness, Recovery and Rebuilding and hosting many community meetings in the aftermath, which helped reduce the negative impact for homeowners.
This guidebook is dedicated to the people of Montecito impacted by the 2018 Montecito Debris Flow and the first responders who provided rescue and aid. We greatly appreciate the personal responses of those in the community who shared their personal experiences. Their stories will provide a better understanding of the perception of the debris flow hazard and response, so that we will be better prepared for future hazards. Critical comments, suggestions, and editing for this guidebook by Joshua Schwartz, Richard Heermance, Phil Hogan, Larry Gurrola, and an anonymous reviewer are appreciated. The research is funded in part by the National Science Foundation Award Number 1830169 (USA), Wildfire Rain (USA), and Montecito residents (USA).
Figures & Tables
From the Islands to the Mountains: A 2020 View of Geologic Excursions in Southern California
This volume includes five geologic field-trip guides in the Los Angeles region associated with the 2020 GSA Cordilleran Section Meeting that was scheduled for May 2020, in Pasadena, California. The guides are organized in a generally counterclockwise order around the Los Angeles Basin. The first guide by Burgette et al. provides new slip rates, age constraints, and observations of the active Sierra Madre fault zone that borders the northern side of the San Gabriel and San Fernando Valleys. The Nourse et al. guide takes a new look at the San Gabriel Mountains from a basement and geomorphologic perspective. Further west, Keller et al. provide one of the first published field-trip guides focused on the 9 January 2018 Montecito debris flows that caused 23 deaths. The volume then moves south to Santa Cruz Island, where Davis et al. provide an updated review of the island’s geology within the California borderlands. The final guide returns to the east, where Platt et al. present the unique geology of Santa Catalina Island with a focus on the subduction-related Catalina Schist.