Simple model calculations of the ice thickness for complementing satellite remote sensing of ice extent

Abstract

Optical satellite remote sensing (MODIS and MERIS) enables to obtain ice coverage imagery during cloud free days. In the mid-latitudes cloud free days occur rather seldom during a single ice season. This reduces the value of optical satellite imagery products meant for operational purposes. In climate research, long time series of ice extent are relevant for providing information on the necessity of ice breaking services. Everyday ice breaking on shipping routes requires information about ice types and ice thickness. While the extent of ice and also some ice types are attainable from satellite remote sensing, a simple ice model can be used for the calculation of ice thickness in the case of fast ice. In this study satellite remote sensing is used for the calculation of ice extent and the identification of ice types in the Gulf of Riga. MODIS reflectance data with spatial resolution of 250 m were used for the analyses of ice cover extent in the period of 2000–2005. The winters of 2000/2001 and 2001/2002 were mild, the maximum ice extent was less than 2000 km2 and the ice season lasted from January till the end of March. The winter of 2002/2003 was quite severe and the Gulf of Riga was fully covered with ice (up to 16000 km2). Pa¨rnu Bay is the area where fast ice forms every year and ice breaking is performed to keep Pa¨rnu harbor operational. A simple combined ice model is therefore applied to Pa¨rnu Bay. Ice modeling is performed when normal and severe winter temperatures occur. The length of ice period is compared with the satellite data from MODIS and visual observations as well. Ice thickness in its turn is compared with the in situ measurements made during these winters. Air temperature, water temperature, wind speed, humidity and solar radiation are the data used in our model calculations. Modeling periods last from December 1 to April 30 in 2002/2003 and 2003/2004 (data from Pa¨rnu Sadam AS). A combined model is used to simulate ice thickness [5]. Initial ice formation and melting when water has access to ice surface is considered to take place both on the top and at the bottom of the existing ice layer. In numerical simulation we have used an equation based on heat fluxes between the atmosphere and the ocean. When fast ice has been formed, it is assumed that water does not have access to the surface of ice any more and ice formation takes place only at the bottom of the existing ice. And therefore an equation based on the difference between water and air temperature is used. Simulated ice thickness agrees well with direct measurements made during severe and normal winters. Simulated snow thickness has been overestimated in comparison with the measurements. The model does not take into account the snow carried away by winds and also the snow that has been compressed (the forming of the so-called snow ice).