Here we use in situ observations to identify spatio-temporal variations of snow particle size in 89 GHz AMSR-E passive microwave satellite imagery. We have correlated high temporal resolution data daily AMSR-E with reference to high spatial resolution Envisat ASAR images to a validation dataset of snow particle size acquired during the Japanese Swedish Antarctic Expedition (JASE) 2007/2008. We have found strong correlations between the 89 GHz AMSR-E data and two different size parameters: particle length and estimated Specific Surface Area (SSA). These correlations have been used to model the grain size variations over the entire region of interest. The daily AMSR-E data have been used to study the evolution of the snowpack over time revealing a seasonal metamorphosis of snow particles at the coast that is largely absent on the polar plateau. Furthermore, the AMSR-E data may exhibit effects from the passing of coastal weather systems on 3-6 day cycles. These effects penetrate to the polar plateau and may represent the drainage of cold air from the plateau drawn-down by passing coastal weather systems.

We have developed a digital image processing method for snow particle size and shape analysis suitable for quick and reliable analysis in the eld. Snow particle size is an important parameter strongly aecting snow cover albedo from seasonally snow covered areas and ice sheets. It is also important in remote sensing analysis because it influences the reflectance and scattering properties of the snow. Alternatively traditional methods based on visual inspection of samples can be used but they do not yield quantitative data. Our method provides an additional alternative to both simpler and more complex methods by providinga tool that limits the subjective eect of the visual analysis and provides a quantitativeparticle size distribution. The method involves image analysis software and field efficient instrumentation in order to develop a complete process-chain easily implemented under field conditions. The results from the analysis are a two dimensional analysis of particle size, shape and distributions for each sample. The developed method improves snow particle analysis being quantitative, reproducible and applicable for dierent types of eld sites.

Snow grain size is an important parameter for determining albedo of the ice sheets and for calibration of optical and microwave remote sensing scattering processes. Snow grain size is a function of the local climate determined by moisture content, air and snow temperature, their gradients within the snow and firn, and wind patterns. Furthermore, it is an indicator on snow metamorphism. We have developed The Digital Grain Size Properties method (DGSP-method) using object oriented image analysis of very high resolution snow grain size images. Commonly used methods are based on visual interpretation, which is a subjective method providing only mean grain size does not retrieve size distribution within each sample.

This is a first attempt to validate satellite information by the in situ measurements from JASE (Japanese Swedish Antarctic Expedition) 2007/2008 using digital image processing. The DSGP-method is based on in-field photography of snow and pixel-based object oriented image analysis. The results show shows decreasing grain size towards the centre of Antarctica and larger grains in the coastal areas. The data used to validate is three different products based on two different types of optic satellite sensors; MERIS (Medium Resolution Imaging Spectrometer) and MODIS (Moderate Resolution Imaging Spectroradiometer).  A first validation captures a cluster relation between grain size in the coastal and at the plateau and optical satellite reflection.

Understanding spatial snow particle size variations are key to help interpretation of remotely sensed data of snow cover. In the case of Antarctica, remote sensing is the only viable option to estimate the surface mass balance of the ice sheet on continental scale. We have investigated snow particle size variability along a transect from the coast onto the polar plateau in Dronning Maud Land, Antarctica, to better understand the spatial and temporal variations in surface snow properties. Two daily samples were collected during a 55 day traverse to capture the regional variability. Local variability was assessed by sampling in grids at selected locations and the particle size and shape distributions for each site was analysed through digital image analysis, which has the benefit of yielding large quantities of reproducible quantitative data without the need for advanced laboratory analysis. The results provide an understanding of the complexity of snow particle size variability at different scales and show a variability range from 0.18–3.31 mm depending on the sample type (surface, grid or pit). We can verify relationships between grain size and both elevation and distance to the coast (moisture source) but have also identified regional seasonal changes, particularly on the lower elevations of the polar plateau. Our data provide possibilities to quantitatively assess the optical properties of the surface snow for remote sensing. The details of the spatial and temporal variations observed in our data provides a basis for further studies of the complex and coupled processes affecting snow particle size and the interpretation of remote sensing of snow covered areas.

Maintaining long time series of observations of the Cryosphere is a key issue in climate research. Long observational time series involve problems due to change in methodology or observers. In order to extend time series and introduce new methods, careful comparisons must be made to ensure homogeneity in the observational data. We have compared an established method for snow grain-size observations used by the Abisko Scientific Research Station (ASRS) in northern Sweden, based on visual interpretation, with a newly developed method for Digital Snow Particle Properties (DSPP) analysis. Transition from subjective visual method into digital reproducible analysis creates less subjective and more comparable results. The ASRS method generates size classifications excluding quantitative analysis size ranges. By determining the sizes of the classified snow using the DSPP-method, actual size ranges for classified snow can be established. By performing a digital analysis of the reference samples and the snow samples classified, we can compare the ASRS classification system to existing official classification systems. The results indicate underestimation of the visual particle size in comparison to the reference samples. Our results show how to quantify the historical data set, which enables us to perform quantitative analysis on the historical data set.