C21D-01 INVITED
Snow cover data records from satellite and conventional measurements
A major goal of snow-related research in the Climate Research Division of Environment Canada is the development of consistent snow cover information from satellite and in situ data sources for climate monitoring and model evaluation. This work involves new satellite algorithm development for reliable mapping of snow water equivalent (SWE), snow cover extent (SCE) and snow cover onset and melt dates, evaluation of existing snow cover products such as the NOAA weekly data set with in situ and satellite data, and the reconstruction and reanalysis of snow cover information from the application of physical snow models, geostatistics and data assimilation methods. In the context of the International Polar Year, a major effort is being made to develop and evaluate snow cover information over the Arctic region with a particular focus on the dynamic spring melt period where positive feedbacks to the climate system are more pronounced. Assessment of the NOAA daily and weekly SCE products with MODIS and QuikSCAT derived datasets identified a systematic late bias of 2-3 weeks in snow-off dates over northern Canada. This bias was not observed over northern Eurasia which suggests that regional differences in variables such as lake fraction and cloud cover are systematically influencing the accuracy of the NOAA product over northern Canada. Considerable progress has been made in deriving passive microwave derived SWE information over sub- Arctic regions of North America where pre-existing algorithms were unable to account for the influence of forest cover and lake ice. Previous uncertainties in retrieving SWE across the boreal forest have been resolved with the combination of 18.7 and 10.7 GHz measurements from the Advanced Microwave Scanning Radiometer (AMSR-E; 2002-present). Full time series development (1978-onwards) remains problematic, however, because 10.7 GHz measurements are not available from the Special Sensor Microwave/Imager (1987-present). Satellite measurements coupled with lake ice model simulations have illustrated frequency dependent, seasonally evolving relationships between brightness temperature and lake fraction across tundra regions. A potential solution based on the temporal evolution of 37 GHz AMSR-E measurements shows some promise as this was found to be significantly correlated with field measurements of tundra SWE, and to be relatively insensitive to lake fraction. New pan-Arctic (N 60°N) snowmelt onset and end date records (2000–2006) were produced from enhanced resolution (4.45 km) QuikSCAT (QSCAT) Ku-band backscatter measurements. The goal is to merge this with melt onset information from other components of the cryosphere (e.g. glaciers, ice caps, ice sheets, lake ice, sea ice) to provide an integrated circumpolar melt onset and duration dataset for climate monitoring and research on cryosphere-climate links and feedbacks. A major challenge is expanding the relatively short time period of Ku-band satellite measurements with historical C-band data (i.e. from ERS-1). Geostatistical methods and snow cover modeling were used to develop a 10-km gridded SWE dataset over Quebec from 1970-2005 for climate studies and evaluation of the performance of the Canadian Regional Climate Model.
C21D-02 INVITED
Progress and Ongoing Issues in the Development of a Passive Microwave Sea Ice Extent/Concentration Climate Data Record
Passive microwave sea ice products provide one of the longest continuous satellite data records with over thirty years of data. The consistent and near-complete data record is the foundation of one of the most dramatic indicators of climate change – the significant long-term decline in Arctic sea ice extent. Thus, passive microwave sea ice products are an important climate record. However, the products are currently lacking key properties to be considered true climate data records (CDRs). These include: (1) minimal metadata that does not current standards, (2) limited error statistics and a lack of grid cell or even granule- level quality assessments, (3) limited intersensor calibration periods to assure the highest consistency possible, and (4) lack of one single authoritative algorithm resulting in several different sea ice estimates. Several strategies are underway to address these deficiencies. First, improved metadata is being developed to meet current and future standards. Second, more detailed quality assessment standards are being researched. Third, more thorough intersensor calibration is being conducted, using longer overlap periods and higher quality sensors for a baseline. Finally, a variety of algorithms are being investigated, as well as data fusion methods, to obtain a single sea ice concentration/extent product that is high quality and consistent over the entire record. Collaborations are underway with U.S. and international groups, including a European project funded through EUMETSAT, to work together to produce CDR-level sea ice products and to coordinate with other cryospheric CDR efforts.
C21D-03
Spatial and temporal trends in Arctic temperature climate data records
The generation of climate data records (CDRs) is a critical step in providing the necessary information for scientists, decision-makers, and stakeholders to make adaptive choices that could improve the nation's resiliency to environmental change and variability, maintain our economic vitality, and improve the safety and comfort of U.S. citizens. These CDRs are particularly needed for the Arctic, where existing evidence suggests climate changes are occurring more rapidly than in most other regions of Earth. In this presentation, I use six temperature data sets over Arctic Ocean areas to examine the following research questions: - Are there significant differences in Arctic temperatures among the data sets? If so, which data sets are statistically different and over what areas? - Are there significant differences in the trend of Arctic temperatures among the data sets? If so, which data sets are statistically different and over what areas? - Do variations in temperatures among the data sets lead to significantly different modeled ice thickness distributions? Discussion of the results will show that there are noticeable differences among all of the datasets, and no two datasets are statistically indistinguishable for all months. Spatially, the largest differences in temperature are centered on areas of the marginal ice zone. For trend analyses, some datasets show warming and some show cooling in the eastern Arctic, and the differences in the data sets lead to significant differences in modeled ice thickness.
C21D-04 INVITED
Greenland ice sheet surface air temperature and accumulation rate reconstruction (1840- 2007) from in-situ data records
Meteorological station and ice core records are combined with regional climate model output to develop a continuous 168-year (1840-2007) spatial reconstruction of seasonal mean Greenland ice sheet near-surface air temperatures and ice sheet snow accumulation rates. Independent observations are used to assess and compensate systematic errors. Uncertainty is quantified using residual non-systematic error. Spatial and temporal temperature variability is investigated on seasonal and annual time scales. We find that volcanic cooling episodes are concentrated in winter and around western Greenland. Warming trends coincide with an absence of major volcanic eruptions. Year 2003 was the only year 1840-2007 with a warm anomaly that exceeds three standard deviations from the 1951-1980 base period. The annual whole ice sheet 1919-1932 warming trend is 33% greater in magnitude than the 1994-2007 warming. The recent warming was, however, stronger along western Greenland in autumn and southern Greenland in winter. Spring trends marked the 1920s warming onset while autumn lead the 1994-2007 warming. In contrast to the 1920s warming, the 1994- 2007 warming has not surpassed the northern hemisphere anomaly. An additional 1.0-1.5°C of annual mean warming would be needed for Greenland to be in phase with the Northern Hemispheric pattern. We thus predict that the ice sheet melt rates and recent mass deficit will continue to grow in the early 21st century as Greenland climate catches up with the Northern Hemispheric warming trend and Arctic climate warms according with climate forecasts. Reconstructed accumulation rates exhibit significant inter-decadal trends. Spatial and temporal surface mass balance is reconstructed from melt intensity derived from the air temperature and accumulation reconstructions. The effect of recent and past century warming on surface mass balance is presented.
C21D-05
Development of microwave satellite-based snow climate data records
Snow cover is an important component of the climate system and is also particularly sensitive to changes in climate. Numerous methods of monitoring snow cover from satellites have been developed. Records of snow cover that use visible satellite products, such as snow cover extent from the NOAA weekly charts, approach or meet the expectations of a long-term climate data record (CDR). Multiple other sources of snow information are available, yet fall far short of the requirements for a CDR. For example, multiple microwave algorithms of snow extent, snow melt, snow depth and snow water equivalent (SWE) have been proposed using past and existing passive microwave satellite instruments (SMMR, SSM/I and AMSR-E). More recently, NOAA has used AMSU sounder data to monitor snow cover. Meanwhile, active microwave instruments, particularly Ku-band and C-band scatterometers, have been shown to be useful for monitoring of snow. The use of multiple instruments, with limited intersensor calibration, and the use of multiple algorithms, limit the development of microwave-based snow CDRs. Furthermore, the existing records often lack the necessary error statistics, ancillary data or metadata necessary for development of CDRs. Possible directions for microwave-based or microwave-enhanced snow CDRs are discussed for Northern Hemisphere seasonal snow cover and for the Greenland ice sheet.
C21D-06
Reconstructing mid- Holocene glacial fluctuations in Glacier Bay, Alaska
The fjords of Glacier Bay National Park and Preserve, Alaska, contain a natural archive of interstadial wood in glacial sediments deposited through the Holocene. The tidewater glaciers of Glacier Bay have retreated over 100 km since the fist recorded observation of their termini in 1794. During this period of retreat and isostatic rebound, hundreds of interstadial in situ tree stumps, woody debris, and organic litter horizons have been uncovered. Radiocarbon analyses of numerous samples, for the current study and by various researchers in the past, have thus far yielded evidence of multiple near-fjord length cycles of advance and retreat throughout the Holocene. Here we focus on the history of tidewater response to the period about 5 ka yrs BP, as it has been identified as a time of rapid climate change by a number of different global climate records, including glacially derived ones. Evidence of this transition, from a warmer early-Holocene to a cooler late-Holocene at ~ 5.2 ka yrs BP, is being discovered in recently deglaciated terrain in the Gulf of Alaska and the western Canadian Cordillera. Recent investigations in the mid-bay area of Glacier Bay have resulted in the finding of interstadial wood samples that cluster in the 4 – 5 14C yrs BP time period. The tree- ring record in Glacier Bay primarily provides a record of past temperature change. Tree-ring cross-dating of these samples and the development of a floating ring-width series, with annual resolution of the mid- Holocene wood analyzed, will allow for the identification and study of decadal to millennial-scale climate variability. By combining radiocarbon and stratigraphic methods with the dendrochronological analyses we are reconstructing the patterns of glacier advance across Glacier Bay during this dynamic period, in turn providing insight on the response of the glacial climate of southeast Alaska during a time of global climate flux.
C21D-07
Climatic Implications of the Mechanical Collapse of the Laurentide Ice Sheet
Although it is well-established that the Laurentide Ice Sheet collapsed over Hudson Bay ca. 8500 14C yrs bp, the role of climate in this event remains ambiguous. We use a finite-element glaciological model to test the hypothesis that the ice sheet would have survived throughout the Holocene via mass-balance feedbacks if its marine-based core had not disintegrated mechanically. Experiments that incorporate a parameterization of the Weertman (1974) marine instability mechanism show a complete Laurentide deglaciation in good accord with geologic data. When the instability is suppressed, the ice sheet contracts from its late Wisconsin maximum, but then stabilizes over Hudson Bay where it remains for the duration of the Holocene. The latter simulation uses a mass-balance scheme based on NCEP2 Reanalysis data; thus, ice- sheet changes are driven by modern environmental conditions. Our results place emphasis on the role of glacio-isostatic feedbacks in glacial terminations. Moreover, the possibility that the Laurentide Ice Sheet could still exist over Canada under present climate raises important questions about what defines a glacial cycle.
C21D-08
A Potentially Non-Steady State Pinedale Glacial Maximum, as Indicated by Half Moon Lake Glacial Valley, Wyoming
The greatest extent of glacial ice during MIS2 (Wisconsinan) in the western US may record a short-lived (sub- millennial) cold event rather than an extended Last Glacial Maximum, based on modeling experiments simulating the Pinedale moraines of Half Moon Lake and adjacent valleys near Pinedale, Wyoming. In some locations including the Half Moon Lake valley, Bull Lake (MIS6) moraines lie well down-valley (2 km) of Pinedale moraines, whereas nearby the moraines are much more closely nested (e.g., Fremont Lake valley, 0.5 km). In a simple flow-line glacier model of Half Moon Lake valley, the subglacial topography (steep upper reaches feeding a nearly flat and locally overdeepened region down-glacier) introduces strong hysteresis behavior with abrupt transitions. We have been unable to find any steady conditions that would grow a steady-state glacier ending at the Pinedale moraines. Instead, the ice preferentially terminates either well up-valley, inside modern Half Moon Lake, or advances to the Bull Lake terminal moraines. In the model, advance of the glacier terminus past Half Moon Lake thickens the ice up-valley of the lake, raising more of the glacier into the accumulation zone and causing further advance. If we specify a warming event as the ice reaches the Pinedale moraines, a steady state Pinedale terminus is possible for a narrow range of parameters; smaller warming allows continuing advance, and larger warming triggers retreat. The modeled time-scale for advance from Half Moon Lake to the Pinedale moraines is typically some centuries for climatic perturbations tested, suggesting the hypothesis that the Pinedale maximum at this site records a short-lived event perhaps linked to the Dansgaard-Oeschger or Heinrich oscillations of the North Atlantic. Simulations for the adjacent Fremont Lake valley, in which the Bull Lake terminated up-valley of any prominent flattening of the valley floor, show more-nearly linear dependence of terminus position on snowline elevation, and faster response of the terminus position to climate change. The moraine positions of the two valleys can be explained if the Bull Lake was similar to or only slightly colder than the Pinedale, but the Bull Lake cold persisted longer.