What is the sensitivity of permafrost to a warming Arctic, and what are the ecological and societal consequences?
Permafrost occurs over vast, often inaccessible regions and underlying physical processes operate over a broad range of spatial and temporal scales. These processes cannot be observed solely with a single in-situ or remote sensor. Long-term satellite observations, intensive field and airborne campaigns and new ways to analyze observational data and integrate them with large-scale Earth system models are all needed.
- Project Permafrost integrates Earth observations from different satellite platforms, field studies, UAS and piloted aircraft to improve how we measure the pace of permafrost degradation caused by a rapidly warming Arctic climate.
- Multi-scale and multi-platform observations will reveal linkages to ecosystem functioning and document changes in growing season in Northern latitudes.
- Project Permafrost combines large and disparate data sets and powerful analytics, serving as an important prototype for driving the Earth Lab analytics platform development.
Primary Contact: Jeffery Thompson
How can we better harmonize across data?
Critical to advancing all Earth Lab endeavors is refining our understanding of data resolution effects on observed phenomena and reducing the uncertainty across scales. Project Data Harmonization examines satellite, UAS, and field-based data collected at varying resolutions to develop metrics characterizing the progression of basic spatial and statistical characteristics such as length, area, density, texture (heterogeneity), and spatial/temporal dependence (autocorrelation). The rates of such variation have never been determined systematically across all resolutions, in spite of wide acknowledgement in soil science, terrain analysis, and hydrologic modeling that such variations can impact physical systems modeling. Part of the challenge is that variations are phenomenon-based, i.e., some data types are more scale-sensitive than others. When data collected at multiple spatial and temporal resolutions are integrated into process models, these sensitivities can introduce bias. When such integration involves social data sets (population density, land use) whose collection is fixed at only a few pre-determined resolutions, spatial and temporal scaling problems are intensified. This is a ‘cross-cutting’ project that will add significantly to the Earth Lab analytics initiative.
Primary Contact: Barbara Buttenfield (Department of Geography)