Fig 1.
Coordinates are in latitude and longitude.
Table 1.
Data description.
Fig 2.
Surface temperature observations for each RWS road status.
Fig 3.
Scatterplot of RWS and mobile surface temperature observations.
Nonparametric (LOESS) prediction line is shown in blue, the 0°C level in solid red, and 10°C in dashed red.
Fig 4.
Scatterplot of RWS surface and air temperature observations.
Nonparametric (LOESS) prediction line is shown in blue, the 0°C level in solid red, and 10°C in dashed red.
Fig 5.
Histograms of observations on RWS (left) and mobile (right) sensors.
Fig 6.
Average difference between each mobile–RWS observation pair.
Stations marked with “R” measure surface temperature with a Vaisala DRS511, embedded in the asphalt, while stations marked with “Opt” have optical Vaisala DST111 instruments. When the RWS and/or the mobile observation is above 8°C, the pair is discarded. Further, each mobile–RWS pair have at least 40 measurements.
Table 2.
Number of distinct sensors per year.
Table 3.
Number of observations by mobile and RWS road status.
Fig 7.
Estimates and 95% confidence intervals of fixed coefficients for the inference models.
Model A2 excluded due to term instability.
Table 4.
Estimates of fixed effect coefficients and quality metrics for the inference models.
Table 5.
Estimates of fixed effects coefficients and performance metrics for the prediction models.
Table 6.
Linear coefficients for adjusting the mobile Teconer RTS411 road surface temperature observations.