Friday, 27 November 2015

BBC and other press coverage of a recent field campaign

A recent field expedition to the Khumbu Glacier by a team of researchers from the University of Leeds and the University of Sheffield was covered by the BBC. The team confirmed satellite observations that several supraglacial ponds down the easterly side of the Khumbu Glacier were coalescing to form a larger glacial lake, similar to other glaciers in the region. The Khumbu is the highest glacier in the world and is used by mountaineers for access to Mount Everest. 

Other press coverage included The Washington Post, Men's Journal, and Atlas Obscura


Khumbu Glacier field campaign - Oct/Nov 2015


A team of researchers from the University of Leeds and the University of Sheffield recently completed a four week field campaign on the Khumbu Glacier in Nepal. The Khumbu Glacier is the highest in the world and every year a small section of the upper glacier becomes the home to Everest Basecamp in Nepal.

Access to the Khumbu valley was by a five day walk with two additional acclimatisation days along the Everest Basecamp trail. Our team camped just off-glacier, a short walk from a small number of trekking lodges at Lobuche. Logistical support and research permissions were organised by Himalayan Research Expeditions. Our guides were invaluable on the glacier and the kitchen team were always ready with hot food on our return! 
Data collection involved Structure-from-Motion ice cliff surveys, GCP georeferencing, and supraglacial pond depth surveys and instrumentation.
Heading along the trail

Our campsite following snowfall
Looking towards the Khumbu Glacier


It is widely known that Himalayan Glaciers in this region are losing mass year on year, though the presence of rocky debris on the surface of glaciers prolongs their response to climate change. The debris cover, which is generally thickest at the terminus of a glacier and becoming thinner at higher elevations, changes the spatial distribution of maximum surface lowering, which occurs where debris is thinner owing to the insulating effect of a thick rock cover. The ablative role of supraglacial ponds and ice cliffs, which are widespread on such glaciers, is little quantified. This is predominantly owing to difficult and hazardous access for collecting field data. Ponds and ice cliffs therefore form the basis of my research on the Khumbu Glacier. 

Ongoing remote sensing analysis from fine-resolution satellite imagery is been used to reveal multi-temporal supraglacial pond dynamics by semi-automatically classifying water bodies. An increasing trend observed on other glaciers in the region is of interest and concern for several reasons. Large glacial lakes forming at the terminus of debris-covered glaciers can pose a potential outburst flood risk in some circumstances, requiring monitoring and remediation efforts to avoid a high-magnitude flood which can travel long distances downstream. Supraglacial water storage also has the potential to mitigate increases in meltwater generated under a warming climate. Ponded water also absorbs incoming solar radiation and this thermal energy is transmitted to the ice below, although this may be through a saturated sediment and debris layer. Exposed ice cliffs often exist adjacent to dynamic ponds and may feature a thin debris layer, reducing their albedo and hence increasing their capacity to melt. Capturing pond and ice cliff dynamics using satellite imagery alone is not possible, owing to revisit times, potential cloud cover and illumination issues, and cost of acquisition. Field access to the features permits surveys and instrumentation to be left in situ to allow continuous monitoring. This is particularly important in supraglacial ponds which exhibit a diurnal thermal regime and can drain englacially, transmitting the stored thermal energy into the glacial interior.

Field monitoring

My field strategy involved repeat Structure-from-Motion (SfM) surveys of ice cliffs, dGPS ground control point identification, and the deployment and retrieval of thermistor strings and pressure transducers in several supraglacial ponds. 

Ice cliffs

SfM is a way of generating fine-resolution 3d models of a surface using photographs from a standard camera which are taken at different positions. The technique was implemented using ground surveys around the ice cliff, although airborne surveys are equally possible and are more time efficient. In this case we did not have access to an aerial platform and helicopter traffic to Everest Basecamp would likely restrict permissions for deployment. A range of cliff sizes, aspects, and locations was captured to allow comparisons of melt rate and morphological evolution. Each survey required a distribution of GCPs around the ice cliff before the photographic survey could be undertaken. GCP markers were distributed and georeferenced with a dGPS on the first ‘lap’ of the ice cliff. Photographs would then be taken during one or two more circuits of the cliff to allow a range of vantage points including high and low viewpoints. GCPs would then be collected on a final circuit. The surface of the dynamic areas of the glacier studied were generally rugged and unstable which limited surveys to two cliffs on a given day.

One of the ice cliffs and ponds surveyed

GCP georeferencing
Velocity measurements of glaciers are generally conducted using remotely sensed imagery. On debris-covered glaciers this can be with optical or radar imagery. Typically the availability of appropriate imagery means velocities below 10 m per year cannot be resolved and these regions are defined as ‘stagnant’. Recently it was shown using fine-resolution imagery from an unmanned aerial vehicle that this categorisation may only loosely be applied, since notable surface motion may still occur. During the Khumbu field campaign I identified a number of boulders distributed in the lower ablation area of the glacier which were georeferenced with a dGPS. A repeat survey in May and October 2016 will reveal both horizontal and vertical displacement, which can be used to validate remotely sensed observations since the precision is far greater (on the order of mm - cm).

Pond surveys

Pond surveys were tailored to assessing water storage dynamics and thermal characteristics. Thermistor strings with temperature loggers at 1 m intervals were used to monitor temperature changes, in addition to a pressure transducer to capture water level change. Most ponds encountered were partially frozen at the start of the field campaign, limiting measurements of depth, which were taken with a plumb line. In May 2016 a robotic surface water vehicle will be deployed with the aim of obtaining fully distributed depth and temperature measurements. 

Conducting a pond survey

Most ponds were frozen on the surface by the end of the campaign, requiring access though up to 10 cm of ice for instrument retrieval.

Instrument retrieval on a frozen pond


Saturday, 5 September 2015

An improved method to represent DEM uncertainty in glacial lake outburst flood propagation using stochastic simulations

Cameron S. Watson, Jonathan Carrivick, Duncan Quincey
August 2015

By C. Scott Watson
My first paper, co-authored with two of my PhD supervisors, was recently published in the Journal of Hydrology. We investigated two commonly used first-past Glacial Lake Outburst Flood (GLOF) assessment techniques (the GIS-based Modified Single-Flow-direction and a 1D hydrodynamic model implemented in HEC-RAS) and proposed an improved GIS-based model, which is distributed along with an example data set in the supplementary information. We found that our new model was least sensitive to  Digital Elevation Model (DEM) choice and provides a means of assessing relative inundation probability for a potential GLOF event. The incorporation of stochastic DEM uncertainty is important when modelling and communicating assessments of an inherently complex process, especially where socio-economic implications exist.

The existing MSF model

From my Masters dissertation I was aware of the widely used MSF model developed by Huggel et al. (2003) as a first-pass GLOF assessment technique. However the major drawback is the requirement of a 'filled' DEM, which when using coarse resolution DEM products such as the ASTER GDEM or SRTM, can lead to a large proportion of the valley bottom being artificially raised to allow continued downstream propagation. This often leads to parallel flow artifacts where a length of the valley has been raised to the same elevation, which limits its use for evaluating lateral inundation extent. Additionally,  assessing the 'likelihood of impact' over long study reaches using the visual output is difficult.

The widely used MSF model applied to a Himalayan catchment:
The new MC-LCP model

To improve downstream flow propagation and negate the requirement of a filled DEM, we proposed and tested a new Monte Carlo Least Cost Path (MC-LCP) model. This model is similarly GIS-based but does not require a filled DEM. In our study we used DEMs processed to balance 'cutting' and 'filling' to remove erroneous peaks and sinks, although this is not a requirement. The model is run iteratively (500 times in this study) and produces a numeric output showing how many times each cell was inundated (i.e considered a least cost path for a potential GLOF flow). This iterative process incorporates a new DEM terrain realisation with each simulation to account for vertical DEM uncertainty and allows relative inundation probability to be evaluated. 

We identify several ways in which our model could be developed further, but suggest in the current form it performs well as a first-pass assessment technique. Although the model has no physical basis, comparison with the geomorphic evidence of the 1985 Dig Tsho GLOF event showed good agreement. Robust first-pass assessments are important to guide physically-based assessments, but may be the only feasible option in regions where fine resolution topographic data are not available for such physically based hydrodynamic modelling.
The MC-LCP applied to a Himalayan catchment:


Sunday, 30 August 2015

Software Carpentry workshop - University of Leeds

August 2015

By C.Scott Watson

Software Carpentry workshops are designed for PhD students and early career researchers wanting to develop programming skills to increase productivity. With a NERC funded workshop running on my doorstep as part of the Advanced Training Short Courses programme, I couldn't miss the opportunity to finally learn some coding.

I'd wanted to develop some Python skills since starting my PhD, and with no prior programming experience, this was the ideal opportunity. The workshop ran over three days, firstly introducing us to the Shell (Bash), before moving onto programming in Python and using version control in Git. The third day gave us the opportunity to work in small groups to put our new skills into practice on a small project. 

Day 1 summary

  • Navigating around directories using Bash and manipulating files
    • This led nicely onto using version control in Bash.
    • Version control is vaguely similar to using tracked changes on a work document but can track whole repositories of files and show the history of changes. It's particularly useful for collaborative projects where several people may be working on developing code.
    • Anyone working on the files can pull these repositories to their own computer, make changes to files and code for example, and then push those files back up to the repository so that other collaborators can access them.
  • Functions of Python relevant to environmental Scientists using IPython notebooks and how to document code so that it can be used by other people, or to remind you how it works months after you wrote it.
    • Leading on from this we the looked at how to build testing functions into the code to make sure it works as desired and produces the correct output.

The IPython notebook is a useful way of writing, testing, and documenting code before finalising it in a script:

Day 2 summary
  • More coding in Python including Matplotlib Basemap, which is useful for everything from making a simple overview map to plotting raster or vector data.
  • Using GitHub, which is the online interface for the version control we had been using in the Shell.
  • Looking at documenting code and choosing an appropriate license so that other can  use and modify the code.

Day 3 summary Our group project aim was to use Python to compare point and gridded rainfall data and we decided to use historical Met Office data, thinking this would be nice and easy to manipulate...only it turned out to be a bit of a nightmare. 

Our first task was using Python to download the text files containing the data for available weather stations using the Met Office URL. We then had to clean these text files to remove weird characters and replace no data values, before conducting some basic analysis and plotting up time series graphs. The mini project certainly put the time demands of apparently simple projects into perspective and got us using our newfound skills, including all collaborating through a Git repository.

The shell:

Plotting up our Met stations in Basemap:

Overall the workshop was a great three days and wasn't overwhelming for someone with no programming experience, but equally it seemed useful for those that already programmed in different languages or were coding on a daily basis as part of their PhD. All the instructors and assistants made sure the workshop ran as smooth as it could and the lessons were hands-on and run interactively on the projected displays, so it was generally easy to follow on your own laptop.