While temperature sensors are getting more and more accurate, uncertainty of air temperature measurement has remained mostly unchanged over the past decades. Where once iconic Stevenson screen shelters dominated the professional meteorological landscape, they are now becoming rarer and slowly replaced by smaller cheaper multi-plate radiation shields and fan-ventilated shelters. Are they still the benchmark of precision air temperature measurement or are upcoming technologies like the helical radiation shield from Barani Design Technologies ready to send them the way of the dinosaurs?
Measuring true air temperature is complicated. AWOS weather stations measure "near surface atmospheric air temperature" at a height of two meters according to World Meteorological Organization (WMO) standards. They usually use sensors calibrated in a liquid bath in adiabatic conditions, while real measurement inside radiation shields and Stevenson screens takes place in anything but adiabatic conditions. In layman's terms, sensor temperature in the real world is never in balance with air temperature, thus measurement error (uncertainty) due to varying sensor construction, sensor reaction time (time constant) and self-heating along with radiative heating and cooling is unaccounted for.
True air temperature
What is "true near surface atmospheric air temperature" is somewhat of a mystery. Before comparing various air temperature sensor systems, one must first understand what one is trying to measure...to understand what true air temperature is.
Like any substance, air is prone to heating and cooling through well know energy flows like solar radiation, infrared radiation, convection, conduction and emissivity. Other sensor related influences include dew condensation, evaporative cooling or phase transitions, direct, diffuse and reflected solar radiation, self-heating and of course, the above mentioned calibration procedures.
What affects real air temperature
First, lets take a look at heating from radiative sources such as the sun and infrared heat radiating from the surroundings, which seem to dominate air temperature error and uncertainty. Even though air is mostly transparent, it is well documented that each of its composing gases has a certain light sensitivity or absorbance spectrum and also emissivity (radiative cooling). This radiative heating of air accounts for the difference between incoming solar radiation from earth's sun and radiation reaching ground level as shown in yellow in the accompanying plot. A familiar example is the absorbance of UV light in the upper atmosphere by ozone molecules.
Just like every other substance, air also has the ability to cool itself by radiating heat away in the form of infra-red radiation. This property known as emissivity is different for every material and color and contributes to the error which effects our measurement quality inside every solar screen and solar radiation shield used to house meteorological air temperature sensors.
The energy balance of our measurement systems effects the temperature our sensors read. In the professional community, it is widely believed that a larger solar screen is more accurate. Why some may think this, we will attempt to explain later, but first lets flip the cards around and look at air temperature from the perspective of an air molecule flowing in the wind two meters above ground according to WMO standards. (For this illustration, it is not important whether it is N₂, O₂, Ar, CO₂, H₂O, O3, NO, NO₂ or any other molecule or dust particle composing air.)
Temperature from the viewpoint of air
On a partly sunny spring day with remnants of intermittent snow cover over grassy fields an air molecule called Caeli moved through the shade of a cloud, where it reached an equilibrium temperature of 15 °C. In the shade, the atmosphere was in balance and Caeli was emitting exactly the same amount of heat through emissivity as was receiving from the surroundings like the ground, vegetation and from diffuse solar radiation while the sun's direct solar radiation remained hidden behind a cloud.
Crossing the threshold of shade into the sun, direct solar radiation bombards Caeli with five times the energy of diffuse radiation. The ground below is of no help since it too is heated by the sun and radiating more heat toward Caeli's bottom than the ground in the shade of a cloud. While exact calculation of Caeli's warmup is beyond this article, her temperature rise is almost immediate and Caeli and her molecular friends find themselves dancing at 3 °C higher temperature than in the shade.
Stevenson vs. Barani
As Caeli's journey continues, she suddenly hits a white gauntlet. Smacking directly into a Stevenson screen shelter, Caeli flips upside down and slides through its slots into a chamber hidden from the sun and filled with thermometers. She quickly shakes off the extra heat accumulated in the sun and peacefully slides past a temperature probe. It remains a mystery what her exact temperature was and if she had enough time to find her new temperature balance before squeezing through the back-side louvers and out into the free air in the sun where she quickly regained her warmth. Was Caeli’s temperature drop measurement error caused by the Stevenson screen?
A few meters later her head starts spinning again as she finds herself skimming past a temperate probe. This time she entered a helical solar shield and before she even noticed being in the shade, she rubbed elbows with the temperature probe. The shield's smaller size and easy air access to the sensor gave Caeli almost no time to shake off accumulated heat from the direct sun as compared with the Stevenson screen.
The future of air temperature measurement
What did the sensors inside the Stevenson screen and Barani’s helical MeteoShield Professional measure?
How was the measured air temperature affected by each shelter?
Which measured temperature is closer to the real atmospheric air temperature?
These are the questions aimed to be answered by the Consultative Committee on Thermometry of BIPM in their 2023 - 2027 roadmap and by a new project being launched this year by EURAMET. The first goal, however, will be to identify all of the components of uncertainty by the International Surface Temperature Initiative or ISTI which will start preparing a joint paper with members of CCT on this topic.
WMO radiation shield comparisons
A more detailed comparison of Barani MeteoShield Professional and the Stevenson screen shelter can be found in a WMO radiation shield comparison study by the Royal Meteorological Institute of Belgium (RMI). “Intercomparison of Shelters in the RMI AWS Network”
A notable study was also performed by METEOMET where multiple radiation shields were compared in a winter alpine setting for the effects of snow and sun reflections. “An experimental method for the evaluation of snow albedo effect on near surface air temperature measurements.”