4. Environmental discriminators for Surface-based and Elevated Thunderstorms


  1. Compare Bureau of Meteorology Temperature soundings to ECMWF Era Interim Temperatures, by calculating the MEAN T differences (Tdiff = Tbom - Tera), at all Era Int pressure levels

  2. Compare Bureau of Meteorology Temperature soundings to ECMWF Era Interim Temperatures, by calculating the MEAN TD differences (TDdiff = TDbom - TDera), at all Era Int pressure levels

  3. Calculate standard deviation of all T differences at all Era Int pressure levels

  4. Calculate standard deviation of all TD differences at all Era Int pressure levels

  5. identify stable layers in BoM sounding data

  6. identify stable layers in Era Int data

  7. Calculate percentage of BoM stable layers NOT captured by Era Int

An Atypical Approach to Discriminating ES from SS

3. Era Int

  • - what it is

  • - how it reanalyses the data (interpolate from model levels to pressure levels therefore smoothing of fields?)

  • - surface and upper air observations – what is ingested and when for Australia

4. BoM Soundings

  • - what it is and the network

  • - how is it is detected

  • - strengths and limitations

2. Era Int Verification with BoM Soundings

  • - proximity GPATS to BoM soundings space and time – why +-1 hour and +-0.25deg

  • - BoM soundings capital cities 2008-2014 with GPATS in proximity – BoM TS soundings

  • - BoM TS soundings mapped to EraInt times (00,06,12,18) if sounding time is within 23-00, 05-06, 11-12, 17-18

  • - BoM TS soundings reduced to Era Int vertical resolution

  • - EraInt soundings at same locations and times

  • - EraInt soundings reduced to BoM TS sounding max elevation

  • - Thermodynamic indices calculated using SHARPpy – Tv corrections in CAPE/CIN calculations

Many thunderstorm climatologies use surface and upper air observations to discriminate ES from SS. Again there is concern too few identified Australian ES events would result from observations-based discriminators. As an alternative to sparse surface and upper air observations, the ECMWF Era Interim Reanalysis (hereafter referred to as EraInt) does have finer spatial resolution (spatial information on a 0.75 degrees x 0.75 degrees grid), however the vertical and temporal resolution is courser ( 37 vertical levels, 6 hour temporal resolution). Like other model-based reanalyses, EraInt contains biases. Bao and Zhang (2013) compared 3000 observed soundings from the Tibetan Plateau, a region where observed soundings are excluded from EraInt reanalysis, thus EraInt are independent of observed soundings. Large moist mean biases through lowest 300hPa depth were found (relative humidity +10 to +25% above observed), along with small low level low level cool mean biases (-0.25 to -1.5 degrees Celsius). To investigate if these same mean biases occur in Australia, observed and EraInt soundings at 2 sites (one coastal and one inland) for one year will be compared.

Despite EraInt reanalysis low level moist and cool biases shown by Bao and Zhang (2013), Allen and Karoly (2013) found EraInt surface mixed-layer (50mb depth) CAPE over Australia compared well to 7 years of observed soundings with only small positive biases, whereas EraInt mixed-layer convective inhibition (CIN) was underestimated compared to observations. CAPE, CIN, and other quantities have been used as discriminator components to derive other ES and SS climatologies (a representation of different discriminators are summarised in Table 1. Allen and Karoly’s results suggest EraInt could be used to calculate the near-storm discriminators in Table 1 and construct a climatology ES and SS events, provided discriminators quantities such as MUCAPE, CIN and θe compare well to observed sounding derived MUCAPE, CIN, θe discriminator quantities. This comparison can be performed using observed and EraInt thunderstorm soundings rather than all soundings to reduce the number of zero instability quantity calculations. For accurate comparison of discriminator quantities, observed soundings will be reduced in vertical resolution to EraInt pressure levels, whilst EraInt soundings will be capped at observed sounding vertical extents. Thunderstorm soundings with GPATS lightning within +- 1 hour and +-0.25 degrees latitude and longitude of the sounding location, could be appropriate spatial and temporal windows given similar choices by Thompson et. al. 2003 (soundings within 40km radius and +-3min of thunderstorm reports).

c c c c c & Colman 1990a & Horgan et. al. 2007 & Brooks et. al. 1994 & Thompson et. al. 2007
climatology type (ES/SS) & ES & Severe ES & Severe SS & Severe ES and SS
near-storm discriminator #1 & surface air lifted pseudoadiabatically to 500hPa, had the same temperature or was warmer than the 850hPa air lifted to 500hPa & Proximity sounding within 3deg (333km) and 3 hours of report & Proximity sounding extends above 300hPa, and within 160km and 1 hour of severe storm report & Inflow Base: Lowest pressure level where CAPE >=100 J/kg and CIN >= -250 J/kg

Inflow Base above ground: ES, at ground: SS
near-storm discriminator #2 & NA & proximity sounding has near surface stable layer (subjectively determined) & Surface CAPE < 150 J/kg & Inflow Top: Highest pressure level where CAPE >=100 J/kg and CIN >= -250 J/kg
near-storm discriminator #3 & NA & A level with MUCAPE>0 exists above stable layer & Surface CAPE > CIN & Contiguous levels between Inflow Base & Top meeting CAPE, CIN criteria
synoptic discriminator #1 & 1. storm report on cold side of surface analysed front & 1. storm report >111km (1deg) on cold side of surface analysed front & NA & NA
synoptic discriminator #2 & 2. storm report location ground temperature, dewpoint temperature and wind is spatially consistent to surrounding stations & NA & NA & NA
synoptic discriminator #3 & 3. The surface air on the warm side of the analysed front must have a higher equivalent potential temperature (θe) than the air on the cold side of the front & NA & NA & NA

Results Era Int Verification with BoM Soundings

  • - CAPE, CIN values poor

  • - ThetaE values are ok

  • - Why CAPE/CIn values poor with sounding examples of:

  • – rain affected BoM not captured in EraInt

  • – missing layers BoM captured in EraInt

  • – EraInt missing BoM warm, cool, dry, moist layers

  • – Confirmed with other studies?

  • - POD EraInt for CAPE and CIN Thresholds - are good

  • - FAR EraInt for CAPE and CIN Thresholds - are good

  • - POD EraInt for CAPE, CIN Eff Inflow discriminator criteria - are good

  • - FAR EraInt for CAPE and CIN Eff Inflow discriminator criteria - RESULTS???

  • - POD EraInt for ThetaE discriminator criteria - RESULTS???

  • - FAR EraInt for ThetaE discriminator criteria - RESULTS???

Old stuff - worth keeping???


  1. Dearchived sounding Melbourne & Giles Airport data 2013 (00, 12Z) from the Bureau of Meteorology

  2. Downloaded ECMWF Era Interim T and RH grib data for a restricted domain over Melbourne & Giles Airport sounding sites (airport lat+-5 degrees lat, lon+-5 deg lon). Converted all grib files to netCDF.

For all Melbourne/Giles Airport BoM soundings in 2013 (n=728/n=361):

  1. To estimate T and RH (all levels) at the sounding site, (a) identified 4 Era Interim grid points surrounding the BoM site, (b) 2D linear interpolation was used at all levels

  2. Era Interim TD using RH and TD approximation formula by Lawrence (2005, equation 8, accurate between -40 degC < T < 50 degC, 1% < RH < 100% ). NOTE - TD(T,RH) formula successfully calculated Equation 8 values in Table 1 of Lawrence 2005.

  3. Tbom and TDbom mapped to Era Int pressure levels by identifying Tbom (TDbom) at levels common to both the BoM and Era levels ie no vertical interpolation

  4. calculating Tdiff = Tbom-Tera, TDdiff = TDbom-TDera at all common eralevs

  5. At 00Z and 12Z for Melbourne (00Z Giles only), calculate MEAN(Tdiff), MEAN(TDdiff), STD(Tdiff), STD(TDdiff) at every common Era Interin level where ’n’ > 100. NOTE - standard deviations were calculated using the n-1 method (assumes that arguments are a sample of the population).