GEUS Dataverse
Not a member yet
699 research outputs found
Sort by
Albedo regression of the Greenland ice sheet
No full textThe project investigates the use of machine learning regression algorithms to estimate
the broadband albedo factor on the Greenland Ice Sheet. The algorithms use data from
the Ocean and Land Color Instrument (OLCI) on the Sentinel 3 Satellite and compare it
to in-situ measurements taken from the PROMICE weather stations.
Three different models and an extension are investigated:
1. LIN: A multiple linear regression model
2. KNAP: A 2nd order polynomial regression function derived by Knap et. al. (1999)
3. DNN: A multiple-perceptron deep neural network.
3.1 SMOOTH: An extension of the DNN using past and future estimations
The LIN and DNN models have an average performance comparable to ESA’s
state-of-the-art process SICE (Kokhanovsky et. al. 2019) developed by GEUS. The
KNAP model proved to be less accurate and ultimately outdated. The main cause of the
performance difference between the KNAP and LIN models is an increase in available
spectral bands due to improved satellite sensors. The DNN model has a very small
performance advantage over the LIN model, while the SMOOTH model makes a more
significant increase in performance by reducing the variance of the DNN model. The
study was done on a limited dataset with approximately 2400 samples from only 10
sites. The available features are 21 spectral narrowbands from the top of the
atmosphere and the sun’s and satellite’s angles. When the LIN model was applied on a
mosaic of Greenland it became clear that it could not handle the effect of angles well.
The DNN model proved to do a much better job at this.
When comparing the models to ESA’s state-of-the-art model SICE , the DNN model has
comparable performance, while the SMOOTH model has a small advantage on
average. Further, the DNN model is less restricted compared to the SICE model, as it
does not build on various assumptions, and hence is applicable to more areas
The firn meltwater retention model intercomparison project (RetMIP): forcing data
No full textThis dataset contains the forcing provided to all models for the firn meltwater Retention Intercomparison Project (RetMIP).
The forcing files consist of 1) three-hourly time series of snow surface temperature, net accumulation (positive for snowfall and vapor deposition and negative for sublimation) and melt and 2) initial profiles for firn density, temperature and liquid water content. These fields are available at four sites located in different climate zones of the Greenland ice sheet: Summit, Dye-2, KAN_U and the firn aquifer site (FA). At the Dye-2 sites, two forcing data files are available: one covering 1998-2015 and another one derived from more recent instruments over the 2016 melt season.
At Summit and Dye-2 (for 1998-2015), we use data from the GC-Net automatic weather stations (AWS), at Dye-2 for 2016 we use the AWS from Samimi et al. (2020), at KAN_U data from the PROMICE AWS, and at FA data from the AWS maintained by IMAU at Utrecht University.
Surface temperature and vapour fluxes were calculated at hourly resolution using the surface energy budget closure from van As (2005), as implemented in Vandecrux et al. (2018). The energy budget used measurements of air temperature, humidity, pressure and net shortwave radiation. Downward longwave radiations was measured at Dye-2 (for 2016), KAN_U and FA but were extracted from the closest cell in HIRHAM regional climate model at Dye-2 (1998-2015) and Summit. Sensible and latent heat fluxes were calculated using the bulk approach and upward longwave radiation was solved iteratively at each time step.
Snowfall was derived from surface height observations available at each station, first with an a-priori surface snow density of 315 kg m-3 (Fausto et al., 2016), then adjusted to multiple snowpit observations available at each site.
Gap-free meteorological fields were obtained using the approach from Vandecrux et al. (2018) which adjusts the output from HIRHAM climate model to match at best the observations and use these adjusted values to fill the gaps in the instrumental records.
More information about the RetMIP protocol and on the recommended boundary conditions can be found in Vandecrux et al. (2020) and on the RetMIP webpage.
The model outputs for the RetMIP are available here:
Vandecrux, Baptiste; Peter L. Langen; Peter Kuipers Munneke; Sebastian Simonsen; Vincent Verjans; C. Max Stevens; Sergey Marchenko; Ward van Pelt; Colin Meyer, 2020, "The firn meltwater retention model intercomparison project (RetMIP): model outputs", https://doi.org/10.22008/FK2/CVPUJL, GEUS Datavers
The Effects of Meltwater, Refreezing and Modelled Grain Size on Snow Albedo: Gaining Knowledge from Observations at Weather Stations and Numerical Modelling
No full textThe extent to which snow and ice surfaces undergo melt is mainly dictated by their absorption of shortwave radiation, and thus the surface albedo is a critical parameter for accurately modelling the surface energy balance and mass balance of the Greenland Ice Sheet (GrIS). Multiple variables affect the surface albedo, including the atmospheric conditions, the solar zenith angle (SZA), and the snowpack’s characteristics.
This thesis uses output from a firn evolution model (Vandecrux et al., in review) to explain 8 years of albedo observations from the Kangerlussuaq Upper (KAN_U) automatic weather station (PROMICE; Ahlstrøm et al., 2008). A parameterised model is developed that differentiates between three snow conditions (aged, refrozen, and wet), and uses four variables as input: the SZA; the Sky-Air Temperature Ratio (SATR); the Liquid Water Content (LWC); and the modelled grain size.
When compared to the observed albedo at KAN_U, the model could explain more than 60% of the observed albedo variation for 2009, 2010, 2011 and 2015, but this fell below 30% for 2014. So far, no reasons for these differences have been uncovered. Each of the three snow conditions performed approximately equally well, and only the lightly cloudy skies were seen to be identifiable worse compared to any other cloud cover. The model was also largely consistent between months, with the exception of October, when the poor fit is attributed to the very short days and consistently high SZAs.
An immediate priority is to test the model against albedo observations from other PROMICE datasets. It is unclear how applicable the model will be to other locations, because it explicitly replicates a rising albedo from the early morning to noon. This was consistently seen in the observational data at KAN_U but contradicts the accepted theory. However, it is possible that the approach used by this parameterisation could be easily modified and tuned to alternative locations
Greenland freshwater runoff
No full text
DEPRECATED
This dataset should not be used.
Please use data found in doi: 10.22008/FK2/XKQVL7.
Greenland land and ice sheet runoff. Assumes subglacial routing with k=1.0.
Code to access and query the database: https://github.com/mankoff/freshwater</a
Supplementary files for: Late Quaternary history of Lammefjorden, north-west Sjælland, Denmark
No full textLammefjorden is a reclaimed fjord in north-west Sjælland, Denmark. Sediment cores from the area were collected to study its development after the last deglaciation, in particular the sea-level history. Late glacial and Early Holocene lake and bog deposits occur below marine deposits. Sparse late glacial fossil assemblages indicate tree-less environments with dwarf-shrub heaths. Early Holocene deposits contain remains of Betula sec. Albae sp. and Pinus sylvestris, which indicate open forests. The wetland flora comprised the calciphilous reed plant Cladium mariscus and the water plant Najas marina. Marine gyttja from basins is characterised by sparse benthic faunas, probably due to high sedimentation rates. In some areas, shell-rich deposits were found, with large shells of Ostrea edulis, indicative of high summer temperatures, high salinity and strong tidal currents. A marine shell dated to 6.7 cal. ka provides a minimum age for the marine transgression of Lammefjorden
Greenland Ice Sheet solid ice discharge from 1986 through last month: Gates
No full textGate locations for ice discharge: doi:10.22008/promice/data/ice_discharge/d/v02.
Update 2022-11-19 (v6) Kangerlussuaq gate moved upstream due to retreat. V6 of this product coincides with V64 (and onward) of the discharge product.
Update: Moon (2018) (NSIDC 0642) metadata added. V4 of this product coincides with V56 (and onward) of the discharge product.
Update: Jakobshavn gate moved inland due to front retreat. V2 of this product coincides with V26 (and onward) of the discharge product
Greenland Ice Sheet solid ice discharge from 1986 through 2017: Gates
No full textGate locations for ice discharg
Kort over dybden til redoxgrænsen i Danmark
No full textKortet viser det bedste estimat over dybden til redoxgrænsen i Danmark i meter under terræn. På basis af ca. 13 000 observationer af farveskift i sedimentprøver er der udviklet en model der beskriver sammenhængen mellem forskellige variable og dybden til redoxgrænsen, der beskriver overgangen fra oxiderede til reducerede jordlag. Til modellering er machine-learningmetoden ’Random Forest’ benyttet, og der er anvendt en rumlig gridopløsning på 100 m. Der er endvidere udviklet en metode til estimering af usikkerheden på kortet, som er beskrevet i de linkede publikationer. Størrelsen på datapakken er 31 MB.
The map shows the depth to the redox interface in metres below ground level. The redox interface describes the transition from oxidized to reduced soil layers. Based on approximately 13,000 observations of colour change in sediment samples, a model has been developed that describes the relationship between various explanatory variables and the depth to the redox boundary. The Random Forest machine learning method has been used for the modelling task, and a spatial grid resolution of 100 meters has been applied. Additionally, a method has been developed to estimate the uncertainty of the redox map, which is described in the related, linked publications. Size of datapackage: 31 MB
Geological map of South and South-West Greenland 1:100 000
No full textThe seamless digital geological map is based on the digitisation and harmonisation of 21 geological map sheets at 1:100 000 scale, originally published by GGU/GEUS between 1966 and 2011. This edition updates and expands the 2019 version, which included 16 sheets, by integrating five additional 1:100 000 sheets and selected information from 1:500 000 scale maps in areas lacking detailed coverage. The map also incorporates a simplified geological interpretation of Bjørneøen and Storeø in Godthåbsfjorden based on detailed mapping by Claus Østergaard (2005). The dataset provides a consistent, seamless geological framework optimised for digital display at the 1:100 000 scale.
One GEUS font is enclosed. In order to symbolise the data correct the font need to be installed on the local machine. (see Readme.txt file)
The maps are delivered in ArcGIS Pro and QGIS formats
Greenland Ice Sheet solid ice discharge from 1986 through 2017: Code
No full textCode for Greenland Ice Sheet solid ice discharge from 1986 through 201