Observations: Relative humidity options

[DamienIrving]

Options for relative humidity might be:

• Conversion of AGCD vapour pressure (available from 1971 onwards)
• Multi-Source Weather (MSWX)
• SILO
• BARRA2
• ERA5
• AWRA

[DamienIrving]

According to Beck et al (2022), the MSWX-past data is basically just ERA5 data that has been interpolated to a finer grid (i.e. there's not an appropriate reference climatology to bias correct against). The advantage of using MSWX-past is that they've gone to the trouble of calculating relatively humidity from the ERA5 outputs (ERA5 doesn't directly output humidity).

If we wanted to use ERA5 and do the calculation ourselves, Beck et al says they calculate it from ERA5 dewpoint and air temperatures according to Vaisala (2013, their Eq. 12):

$RH = 100 × 10^{m((T_d / (T_d + T_n)) − (T / (T + T_n))}$

where RH (%) is the relative humidity, $T$(K) is ERA5 air temperature, $T_d$(K) is ERA5 dewpoint temperature, and $m$ and $T_n$ are constants of 7.59138 and 240.7263 K, respectively.

Vaisala, 2013: Humidity conversion formulas. Vaisala Tech. Rep., 17 pp., https://www.vaisala.com/en/lp/make-your-job-easier-humidity-conversion-formulas.

[DamienIrving]

Conversion of AGCD vapour pressure ($e$) to relative humidity is defined by: $RH = (e/e_s) * 100$

The saturation vapour pressure ($e_s$) could be calculated using the AGCD temperature data (if using Python, via metpy.calc.saturation_vapor_pressure(temperature)).

The AGCD dataset provides two vapour pressure values each day: 9am (h09) and 3pm (h15). If we want a daily mean relative humidity value, we'd possibly need to use the daily minimum temperature data to calculate the 9am saturation vapour pressure and daily maximum temperature to calculate the 3pm saturation vapour pressure. The daily mean vapour pressure (and saturation vapour pressure) might then just be the mean of the 9am and 3pm values.

[DamienIrving]

The SILO dataset has daily vapour pressure values as well as derived "relative humidity at the time of maximum temperature" and "relative humidity at the time of minimum temperature" variables. The variables page says that the relative humidity values were "calculated using the vapour pressure measured at 9am, and the saturation vapour pressure computed using either the maximum or minimum temperature (Jeffrey et al 2001)"

[DamienIrving]

The daily vapour pressure data that is produced for input into the AWRA model basically solves the AGCD 9am / 3pm problem. According to the AWRA v7 technical report, they use a weighted average as the input to the AWRA model to reflect the average daily value. The weight is a new parameter for AWRA v7 with an optimum value found to be 0.2 for the 9:00 am and 0.8 for the 3:00 pm values.

If we want to calculate relative humidity from vapour pressure we'll also need the daily average temperature 𝑇𝑎 in order to calculate the saturation vapour pressure. When calculating saturation vapour pressure and ultimately potential evaporation the AWRA system takes the weighted mean of the daily maximum (𝑇𝑚𝑎𝑥) and minimum (𝑇𝑚𝑖𝑛) temperatures with the weights 0.85 and 0.15 respectively. (Unless Tmin > Tmax, in which case Ta = Tmax.)

(We'd just need to make sure of the correct time alignment between variables according to #27)

[DamienIrving]

A rough comparison of the various options can be found in the following notebook:
https://github.com/AusClimateService/npcp/blob/master/obs-comparisons/relative_humidity.ipynb