Fig. 1
Geography map of Lake Orajärvi in Sodankylä. The symbols in the right panel: ○: regular snow and ice thicknesses measurement site; □: SIMB site and, *: Sodankylä weather station.
Fig. 2
Measurements of mean snow (red) and ice thickness (blue), and freeboard (black) on Lake Orajärvi for: (a) 2009/2010; (b) 2010/2011; and (c) 2011/2012 winter seasons.
Fig. 3
(a) The SIMB measured snow and ice temperature fields. The number of sensors counts downward from surface to bottom. The snow/ice interface is marked as a black line. (b) SIMB data derived snow and ice thicknesses applying snow/ice interface as zero reference level. The symbols are snow and ice thicknesses measured near the SIMB site (▴) and at regular measurement sites (○).
Table 1. Calculation of freeboard in terms of various snow/ice compositions present in lake
;; Ice/water; Snow, ice/water;;
;Snow, flooding slush and ice/water;;
;Snow, snow–ice and ice/water;;
;Snow, flooding slush, snow–ice and ice/water
Table 2. The seasonal (Nov.–Apr.) mean values of observed weather data as well as Bias, Root Mean Square Error (RMSE) and correlation coefficient (R) between hourly time series of HIRLAM forecasts and in situ observed wind, temperature, humidity, downward shortwave and longwave radiation and precipitation for winters 2009/2010; 2010/2011; and 2011/2012
09/10Mean (Nov.–Apr.)2.3−9.28642250Sum 192Sum168
10/112.3−10.48546237(Nov.– Apr.)131(Nov.– Apr.)110
11/122.3−6.88741261271218
09/10Bias (Cal –Ob.)1.1−0.212.8−9.8−6.3(Nov.–Apr.)99(Nov.–Apr.)67
10/110.5−0.37.3−8.6−2.2184.8
11/120.8−0.16.4−8.1−4.5325.3
09/10RMSE1.73.418.269.324.410/111.23.714.274.522.611/121.33.513.467.530.209/10Corr.0.780.940.290.740.89
h0.64
d0.83
h0.66
d0.8510/11Coeff.0.770.930.600.740.880.450.680.510.7411/12R0.780.920.480.730.810.580.820.590.82
Fig. 4
The monthly mean differences between HIRLAM forecasts and weather station observations for wind speed (V a ), air temperature (T a ), relative humidity (Rh), total precipitation (PrecT), snow precipitation (PrecS), as well as shortwave (Q s ) and longwave (Q l ) radiative fluxes for winters 2009/2010, 2010/2011 and 2011/2012. Note the PrecT has the same y-axis [−20, 40], but for May of 2009/2010, the difference was 127 mm.
Table 3. The names of HIGHTSI modelling experiments
with snow
local in situ observation0910SL1011SL1112SLHIRLAM forecasts0910SH1011SH1112SHno snowlocal in situ observation0910L1011L1112LHIRLAM forecasts0910H1011H1112H
Fig. 5
The HIGHTSI (experiments SL, c.f. Table 4) modelled (red line) and MODIS observed LIST (blue dots) for winter (a) 2009/2010; (b) 2010/2011; and (c) 2011/2012. The SIMB measured surface temperature is given as the green line in (c).
Fig. 6
The comparison of surface temperature: (a) MODIS versus HIGHTSI (SL experiments, c.f. Table 4); (b) MODIS versus SIMB; and (c) HIGHTSI versus SIMB for winter 2011/2012.
Table 4. Mean Bias Error (MBE), standard deviation (std), Root Mean Square Error (RMSE) and correlation coefficient (R) of HIGHTSI (experiments SL, c.f. Table 3) modelled surface temperature, as well as SIMB and MODIS observations
2009/2010HIGHTSI–MODIS2.94.55.40.932772010/2011HIGHTSI–MODIS2.85.15.80.894162011/2012HIGHTSI–MODIS3.94.96.50.88375MODIS–SIMB−3.14.65.50.9253HIGHTSI–SIMB0.032.83.80.91757
Table 5. HIGHTSI (experiments SL, c.f. Table 3) modelled monthly mean surface heat fluxes for three winter seasons. Q snet : net shortwave radiative flux at surface; Q lnet : net longwave radiative flux; Q h : sensible heat flux; Q le : latent heat flux; F c : surface conductive heat flux; F net : net surface heat flux, that is, the sum of Q snet , Q lnet , Q h , Q le and F c . All fluxes are positive toward surface (heat gain)
YearMonthW/m2
09/10110.1−5.71.1−0.66.92.0120.01−19.42.2−0.0121.44.2010.12−17.91.9−0.919.52.8022.2−17.9−4.2−1.716.8−4.8039.0−35.05.6−1.813.9−8.30411.4−29.610.4−3.213.32.20523.9−7.312.44.49.943.310/11110.2−29.20.1−0.640.211.0120.01−23.8−1.6−1.034.38.0010.16−19.11.8−0.0221.14.0022.3−22.8−1.6−0.918.1−5.1039.3−43.714.7−1.713.4−8.00413.5−31.212.7−1.812.45.70526.9−17.418.23.97.830.911/12110.4−15.93.92.29.1−0.3120.01−11.95.21.18.83.3010.26−10.6−1.1−1.610.4−2.5022.5−15.9−1.0−1.310.6−5.00311.2−34.69.30.88.6−6.30420.9−30.65.9−2.47.10.90525.9−27.324.7−0.17.330.5
Fig. 7
HIGHTSI modelled snow thickness (upper lines), ice thicknesses (lower lines) and freeboard (middle lines) compared to in situ observations for 2009/10 (a), 2010/11 (b) and 2011/12 (c). The zero reference level is the snow–ice interface. The model runs were based on atmospheric forcing from weather station data (red lines) and HIRLAM forecasts (blue lines). The observations are given as black circles and their standard deviations are marked as vertical bars. The observed ice break-up dates of Lake Unari are indicated by red circles.
Table 6. The seasonal maximum ice thickness (H imax ), ice bottom growth and percentage of snow to ice transformation for SL and SH experiments and H imax for L and H experiments for three ice seasons
SLH imax (cm)668157SH758955SLColumnar ice (cm)365627SH396731SLGranular ice (%)463053SH482444LH imax (cm)12013296H12313599
Fig. 8
The HIGHTSI modelled ice bottom growth for (a) 0910SL (red) and 0910SH (blue); (b) 1011SL (red) and 1011SH (blue); and (c) 1112SL (red) and 1112SH (blue).
Fig. 9
Ice core sample collected on 12 April 2012 when the SIMB was recovered.
Fig. 10
HIGHTSI (SL experiments, c.f. Table 3) modelled snow and ice temperatures for winter seasons 2009/2010, 2010/2011, and 2011/2012. For the sake of clarity, snow and ice temperatures have the same vertical scales but different temperature grey scales.
Fig. 11
The comparison of 1112SL modelled (red line) and SIMB observed (dot-blue line) vertical temperature profiles within snow and ice at selected UTC time steps. A normalized coordinate, that is, height/(h s +H i ) is used in the y-axis (0 is surface and 1 is ice bottom).
Fig. 12
HIGHTSI modelled ice evolution without taking snow into account, using in situ weather station data (solid line) and HIRLAM forecasts (dotted line) as external forcing. The winter seasons are 2009/2010 (red), 2010/2011 (black), and 2011/2012 (blue).
Fig. 13
The differences of modelled ice between model runs (without snow: solid lines; with snow: dotted line) using in situ weather station data and HIRLAM forecasts as external forcing. (a) 2009/2010; (b) 2010/2011; and (c) 2011/2012.