NWT RESEARCH DATABASE

Tags: physical sciences, permafrost, hydrology, glaciology, prediction models, remote sensing

Objective(s): To develop a suite of models for predicting the response of discontinuous permafrost in the Hay River Lowland to climate warming and human disturbance from oil and gas exploration, and the consequent change in landcover and river flow regime.

Project Description: The long-term goal of this research project is to develop a suite of models for predicting the response of discontinuous permafrost in the Hay River Lowland to climate warming and human disturbance from oil and gas exploration, and the consequent change in land cover and river flow regime. This will be achieved by meeting the following short-term objectives: 1) map the spatial distribution of permafrost and its change over the past 60 years using aerial photography and satellite images, 2) develop conceptual and mathematical models of hydrological processes, 3) develop a new model of permafrost to simulate its response to climate warming and human disturbances, and 4) couple the hydrological model with the permafrost model to predict the spatial distribution of permafrost and the river flow regime under possible scenarios of climate warming and human disturbance. The Scotty Creek research basin, 50 km south of Fort Simpson, NWT is the primary site for process studies and model development. The adjacent Jean-Marie River basin will be used to test the transferability of models within the Hay River Lowland; however no field work will be conducted in Jean-Marie. Boreal peatlands in the Hay River Lowland offer a unique opportunity to monitor permafrost thaw rates because thaw in peatlands causes a dramatic change in land cover that is easily detectible from ground observation and aerial and satellite imagery. High-resolution (2.5 m), multispectral SPOT5 images will be acquired for the Scotty Creek and Jean-Marie basins in August 2011 and the areal extent of permafrost plateaus will be delineated by supervised classification. Airborne light detection and ranging (LiDAR) will be used to obtain detailed ground surface elevation and canopy height data for Scotty Creek in August 2011. High-resolution (4 m) IKONOS satellite images from August 2000, and seven sets of aerial photographs taken between 1947 and 2008 have been obtained for Scotty Creek. Using the time series of images, sequential maps will be developed that show changes in the spatial distribution of permafrost plateaus, bogs, and fens; and the rate of reduction in the area underlain by permafrost. The unique challenges of this project are in scaling-up from an individual plateau to an aggregate of plateaus, characterizing the storage and routing of water through the network of bogs and fens, and understanding the role of groundwater. It is impractical to run the computationally intensive SFASH model for numerous plateaus. Instead it will be used to define the relationship between the hydrological characteristics and simple geometric parameters of plateaus. Hydrological characteristics refer to the relationship between the amount of water stored on a plateau and the rate and timing of its release. Preliminary analyses suggests that the area (A), perimeter (P), and equivalent hydraulic radius RH (= 2A/P) have a strong influence on the amount and timing of runoff generation from individual plateaus. All of these parameters can be easily measured from the classified images. The statistical distribution of A, P, and RH in a grid cell will be used to simulate the composite runoff generation from all plateaus within the cell. Water-level recorders will be installed along several seismic cutlines to monitor the propagation of storm pulses along their length and throughout the seismic grid. The water storage capacity of bogs and the connectivity between bogs and fens changes dynamically with water level. Using the high-resolution (1 m horizontal, 0.2 m vertical), LiDAR-derived digital elevation model, the water storage, wetland connectivity, and hydraulic gradients between adjacent wetlands for different water levels will be defined. Using the hydraulic roughness and connectivity, a water-level dependent hydraulic transfer function between grid cells will be developed. An area-average storage function for a grid cell will be defined, based on the statistical distribution of wetland size, geometry, and connectivity indices extracted from the classified image. Using the algorithms developed, the hydrological response of the collection of all plateaus and wetlands in a single grid cell to meteorological forcing will be described. Scotty Creek will be divided into grid cells and a basin hydrological model will be built to simulate the cell-average wetland water storage, soil moisture and frost table depth on plateaus, flow between cells, and the stream hydrograph at the gauging stations. The soil-vegetation-atmosphere transfer SVAT algorithm from an existing permafrost model will be used with the physical parameters, such as soil hydraulic parameters and vegetation density, measured in the field or derived from airborne and satellite images. The basin hydrological model will be calibrated using field data from Scotty Creek during 2011-2013, and validated using 2006-2008 data. The model will also be set up for Jean-Marie River for the same time period to examine transferability of calibrated model parameters between the two basins. After calibration and validation, the model will be used to simulate the hydrograph of the Jean-Marie River for the 1970s, 1980s and 1990s using the land cover distribution developed from archived photographs and satellite images. Model simulations will be compared with measured hydrographs to examine if the model is able to reproduce the observed trend of increasing discharge in recent years. Field observations will focus on the receding edges of plateaus. A new 20-m measurement transect perpendicular to the edge of a plateau will be added in the early summer of 2010. Meteorological sensors will be placed in an open (i.e. treeless) bog and below a tree canopy to provide reference data. Several arrays of soil moisture and temperature sensors and ground heat flux sensors will be installed, as well as groundwater monitoring wells along the transect. A similar transect perpendicular to an existing seismic cutline will be installed to examine the effects of linear disturbances on permafrost. Piezometers ranging in depth from 1 m (surface peat) to 4 m (glacial deposit) will be installed in an isolated bog and a fen to examine the dynamics of groundwater flow under recharge (bog) and discharge areas (fen). The Northern Ecosystem Soil Temperature (NEST) model will be modified to simulate discontinuous permafrost. Three NEST models will be set-up in a single grid cell (representing plateau, bog and fen surfaces) and run simultaneously so as to include lateral heat transfer among them. Using this method, the model will simulate historical changes in the relative proportion of plateau, bog, and fen, and the permafrost thickness within each land cover type. The model will be run for all grid cells within the Scotty basin to simulate basin-scale permafrost thaw. To test the model, the 1947 permafrost map will be used as the initial condition. The simulated permafrost thaw will be compared with the observed changes in permafrost coverage from the aerial/satellite documentation, including the latest map based on the satellite images from 2010. Model output will also be compared with the temperature profiles measured by the GSC at four 10-20 m deep boreholes within or adjacent to the study area. Guidance of project will be sought through public consultation with community groups at Jean-Marie and Fort Simpson. The researchers wish to meet with Liddli Kue First Nations, as they have a harvesters group who might be interested in the information being generated. Annual reports to the Aurora Research Institute, and publications will be sent to the communities each year. The fieldwork for this study will be conducted from January 1, 2011 to December 31, 2011.