Video of this trip is posted on YouTube at http://youtu.be/_A2E7FUXXaY. A group of geologists made a cross-country trek to examine the landslide that spawned the large May 31, 2013 Middle Fork Nooksack River debris flow (several stories earlier this summer in this blog). The party consisted of Jon Riedel (North Cascades National Park glacial geologist), geomorphologist Scott Linneman (WWU Geology Department), John Scurlock (MBVRC and photographer extraordinaire) and Dave Tucker (WWU and MBVRC). The purpose was to pinpoint the source of the landslide and examine the two-month-old deposits in the uppermost part of the deglaciated valley below the terminus of the Deming Glacier. We were also curious to determine whether buried ice had played a role in the landslide, and if any bedrock was involved.
The route began at the top of the Rankin Creek Road (grown over but passable to a Forester if willing to sustain some brush scrapes to the paint), and immediately entered old growth on a steep side hill. We passed the top of the slope that slid into the river and partially dammed it on June 6, then descended forested moraine steeply to just above the river. After a brief but nasty battle against twisted willow and alder, we emerged onto the May 31 debris that blankets the valley floor, and it was smooth sailing the remaining mile to the glacier terminus. The trek up the debris-covered valley floor was surreal. The clay in the debris flow had set up like concrete, studded with stones and even large boulders. The brush was covered and shredded, but it had survived and was again leafing out. We felt we were in a disaster zone, wondering when we would find the destroyed city around the next bend. Total distance is only 2.5 miles but it took most of the day.
The landslide occurred in glacial moraine just above the terminus of the Deming
Glacier on the north side of the valley, elevation around 4500 feet (1370 m).
The slope that failed is very young moraine banked (‘inset’) against higher Little Ice Age moraines on the steep slope above the current terminus of the glacier. Comparing the location with the best available USGS topographic map sheet (Baker Pass, Washington 1:24,000, 1989), the landslide source was at the glacier margin when the map’s aerial photos were compiled (either 1974 or 1985), and the moraine may be only that old. We could not reach the landslide source (more on why below), so a volume calculation is necessarily rough at this stage. Using measurements taken from Google Earth as a crude approximation, the landslide involved an area about 120 m wide (parallel to the slope) by 70 m (up and down the slope). Photos suggest the thickness of the landslide may have been as little as 10 m. So, a first order volume approximation is 84 x 103 meters- pretty small to have caused all the devastation in the valley for miles below the source! We’ll discuss that below.
No bedrock component is evident. We saw no buried ice in the moraine on the right bank.
It is possible that the collapsed moraine was saturated by melting snow at the end of May. It was typically unconsolidated glacial rubble and likely gravitationally unstable. Rapid lowering of the glacier’s surface over the past decades removed support from the inner margin of the moraine. At the end of May, a small portion of the moraine slope collapsed, and may have crossed the extreme NW corner of the glacier before it then pitched over a vertical 30-m-high bedrock wall that the glacier is perched on. The flow landing on avalanche snow accumulated beside the river. The snow is preserved even now, late into a hot summer, buried under 1-3 meters of debris flow rubble. The collapse was clearly a debris flow, with at least a small proportion of entrained water, by the time it rode across this snow, rather than a water-poor landslide. If the moraine was truly saturated, the collapse may have transformed into a debris flow even before it reached the bottom of the slope and over the glacier, after a fall of only a few meters. The buried snow near the valley floor just below the rock wall is clearly seen in this photo from John Scurlock: http://www.pbase.com/nolock/image/151726463 The debris flow deposit is remarkably thin here- there was little for it to adhere to on this surface, so little was deposited and the bulk of the debris flow could continue to race down the valley. The debris flow entered the river after crossing the avalanche snow. Terraces remain marking the high point of the debris flow on the left (south) bank, incised and nearly removed in the intervening months..
The debris flow beyond the glacier
The debris flow ran down the river bed, helped along by a river that was full from melting snow. The flow bulked up with alluvium and fine-grained sediment eroded from till-mantled river banks as it flowed down the valley, gaining volume as it went. Much material was left behind in the upper valley, but the flow then entered a narrow, debris-choked reach about 1 mile from the terminus. Either side of the river channel here is mantled with a 20-m-thick debris flow dating to 1927, or landslide debris from a steep undercut slope. (This same slope would fail on June 6, sending a landslide into the river and partially damming it. A secondary debris flow went down the Middle Fork nearly to the Ridley Creek trail crossing). The May 31 debris flow reached the wide channel below the Ridley Creek trail, where the gradient decreased and the flow began to lose momentum. It dropped much of its coarse bouldery load, and by the time it reached the Elbow Lake trail crossing, had become a muddy hyperconcentrated flow.
The Deming Glacier has been receding at a rapid pace since the 1970s. Some information on the glacier’s balance is available in a study by Mauri Pelto:
The northern half of the glacier’s terminus is now sitting on top of a vertical wall of nearly-crackless silvery gray, polished Nooksack Formation sandstone. A central tongue of ice descends the rock, reaching nearly to the valley bottom. The Middle Fork Nooksack gushes full-blown and turbid out of the ice- the river bursts out of the ice in deep bedrock channels. The bottom edge of the ice tongue is undercut by melting and is very difficult to climb onto- access to the glacier is now extremely difficult and would require technical ice and rock climbing. Probably the best time to access the glacier surface is late spring to early summer when there may still be snow piled against the rock face and covering the tongue. The southern margin of the ice extends beyond the rock face almost to the river bank beyond the rock wall, and is covered by steep, unstable rubble.
Future landslide/debris flow hazard
The source of this rubble is a very steep, threatening wall of 20th C moraine that has obviously collapsed very recently, but prior to May 31- the debris flow mantles the lower-most portion of the landslide toe at the bottom of the rubble wall. The freshly exposed moraine wall results from recent landsliding and is at least 30 m high, and is the most worrisome feature we saw in the headwaters. The undated landslide from this wall reached the river immediately at the base of the rock wall. This may have generated a debris flow, hard to imagine it did not. Perhaps a check of turbidity records in the past year or so would reveal it, as it did the May 31 and June 6 events this year. There is a lot more moraine still hanging on the left bank buttressing a steep slope of older moraine in the same way as the May 31 landslide across the valley. However, on the left wall, ice is inset against the moraine. As it melts, the rubble wall will likely become even more unstable and could collapse in a larger landslide than occurred May 31, directly into the river. The ensuing debris flow could be much larger than the May 31 event in the Middle Fork Nooksack. Other than this area, there is plenty of inset moraine all along the valley. Any of this could collapse, particularly along the margins of the thinning glacier as it continues to recede. This place bears watching, not to mention that it is also difficult and dangerous to visit on foot.
Much data has now been gathered on the May 31 landslide and debris flow, and volume and velocity estimates will soon be available.