BAMOS November 2025
20 Article
of persistent rain or drizzle with cloudy and windy conditions, when it is expected that VTD ≈ 0 ° C at all stations. On a monthly and long-term basis, model fitting and‘ signature’ checks are performed. Some interesting examples are described below.
b) Examples of accuracy checking
The monthly average of mean sea level pressure typically has a smooth and weak gradient over the Mesonet footprint. Using relevant BOM stations, an accurate reference surface is obtained against which Mesonet station discrepancies are determined. Since the original calibration, only one station has had a discrepancy exceeding 1 hPa.
The monthly mean maximum temperature for South Australian BOM stations is modelled by regression based on altitude, latitude and the log of distance from the coast( following Marsh, 1987). For July 2021, using relevant BOM stations, jackknife regression residuals were mostly ≤ 0.3 ° C. Using predictors and coefficients based on BOM stations alone, Mesonet residuals were ≤ 0.5 ° C, except for Saddleworth, which was almost + 1 ° C. Since the sensor replacement, this station is no longer an outlier. It is to be noted that because these monthly checks reference BOM data, they also serve as drift checks.
Monthly histograms of VTD reveal characteristic signatures. For January 2021, histograms for a group of stations in the Riverland & Mallee region highlight Alawoona station as atypical( Figure 3). This technique is automated with the Kolmogorov-Smirnov test. The techniques for monthly checks are extended to longterm datasets.
The Mesonet data quality is reported to be similar to BOM’ s( Bridge et al., 2024a). A further study( Bridge et al., 2024b) compared ACCESS model performance with and without Mesonet input and demonstrated increased forecast accuracy within Mesonet ' s footprint.
Capturing micro- and meso-meteorological features
Users interpret the Mesonet data to inform their own short-term forecasts. Common examples include timing the progress of a cool change or a sea breeze. Additionally, all farmers have their own experience-based local models, such as the movement of dry lines or modulations of wind elsewhere in the vicinity.
Inversions ≥ 3 ° C are often observed outside of summer. During these events, wind speed may exceed 3 m / s( Figure 4). In agriculture, a common rule of thumb( broken in this case) is that inversions do not occur when wind speeds ≥ 11 km / h( 3 m / s). Occasionally, wind speeds at 2 m exceed the speeds at 10 m, and / or the directional shear is 30 to 40 ° or more( not shown).
During situations conducive to inversions, topography and landscape dominate the temperature and wind fields. Figure 5 shows the resulting spatial variability in wind direction in such a situation. On the other hand, in a synoptically dominated situation, the Mesonet’ s density facilitates the positioning of mesoscale features such as a low and front( Figure 5). The
Figure 3: Histograms of VTD for January 2021. The group histogram of Riverland and Mallee stations( left) has a tall, sharp peak at ~ 0 ° C with a secondary peak at ~-1.7 ° C. Alawoona histogram( right) has no secondary peak, thus indicating a fault, which transpired to be a kink in the thermocouple cabling. Credit: Warwick Grace
Marsh, L. 1987. The dependence of mean midsummer maximum temperature on position: Some deductions from climatological data. Meteorological Note 177, Bureau of Meteorology.