The type and character of precipitation reaching earth's surface mainly depend on the thermodynamic state of the troposphere. Numerous thermal parameters are used to determine precipitation type, among them the depth of the thermal layer, freezing height, inversion levels, critical temperatures of isobaric surfaces, etc. The study focuses on relations between thickness and temperature parameters above the canopy layer and the precipitation type. The study period covers 1997-2006 cold seasons (November-March). The parameters used in the study are: 1000-500 hPa integrated thickness (ST1000500) and partial thickness ST1000850, ST850700, ST850500, ST700500; also 925 hPa, 850 hPa, 700 hPa, 500 hPa level temperature (T925, T850, T700, T500). Precipitation type data for Lithuanian territory were compiled using the network of the National Weather Board (LHMT) first-order stations (Biržai, Kaunas, Klaipėda, Varėna, and Vilnius). Free atmosphere data were obtained from NCeP 2 database (http://nomad3.ncep.noaa.gov/ncep-data/) and the Kaunas radio-sounding station. The threshold values of the thermal parameters were calculated using one standard deviation (SD) statistical model. If the derived thickness or temperature fall into the ±SD interval, then the value is classified as varying around "normal" ("normal" pertains to statistical representation) (Heppner, 1992). All precipitation types were grouped into three categories: snow (snow, ice pellets), rain (rain, drizzle) and winter mix (snow with rain, sleet). The first part of the paper presents typical values of thickness and temperature for different precipitation types and their spatial and temporal variability. Integrated thickness (ST1000500) near the coastline seems to be always about 23 gpm thinner for snow pattern than in the rest of Lithuania. For rain, ST1000500 appears to be almost constant in the whole territory. However, thickness parameters show a large spatial and temporal variation. The mean values of integrated thickness related to snow type patterns are highest in late autumn, later they start to drop and in early spring reach the lowest values. There is no such seasonal trend for rain patterns: thickness values are also highest in late autumn, but in mid-winter they become minimal, while at the end of winter again begin to rise. A similar trend was observed for typical temperatures at the standard layers. The threshold values of the thermal parameters are presented in the second part of the paper. These values should be used in the diagnosis of precipitation type. The best indices in the diagnosis appear to be ST1000500 and ST1000850. These parameters indicate the correct number of precipitation type events in 82% of show cases and in 85% of rain cases. Snow is most likely to occur when ST1000500 < 5285 gpm and ST1000850 < 1285 gpm, while rain is most probable at ST1000500 > 5285 gpm and ST1000850 > 1295 gpm. The best temperature indices appear to be at the 925 hPa level (T925).… [ABSTRACT FROM AUTHOR]