|
||||||||||||||||||||||||||||||||||||||||||
|
Analyses & Numerical Prediction Analysis and Prediction Operations Bulletin 54
|
The Bureau of Meteorology Research Centre/Australian Integrated Forecast System (BMRC/AIFS) Model Output Forecast (MOF) system is an application designed to generate forecast guidance based on Model Output Statistics (MOS).
MOF (Version 1) is the operational system in the NMOC implemented on 27 August 1998. The MOF guidance is derived from LAPS375 (00UTC and 12UTC) model output twice per day.
Analysis and Prediction Operations Bulletin No. 43
Derivation of Two Day Model Output Statistics Coefficients
MOF (Version 2) is the operational system in the NMOC implemented on 25 November 1999. The MOF guidance is derived from GASP (00UTC and 12UTC) model output twice per day.
Analysis and Prediction Operations Bulletin No. 50
Derivation of Seven Day Model Output Statistics Coefficients
MOF (Version 3) provides a re-derivation of all the MOS coefficients for GASP and LAPS375 (00UTC and 12UTC) and is intended to replace both MOF (Version 1) and MOF (Version 2).
The MOS coefficients for both GASP and LAPS375 (00UTC and 12UTC) models have been generated using a database of model output and observations covering the period 1 July 1996 to 30 April 2001.
MOF (Version 3) guidance is derived from the GASP and LAPS375 model output for 621 stations out to seven days from model run time, and includes all current Regional Forecasting Centre (RFC) public forecast sites. (See Figure 2 for the distribution of sites.) The full range of daily forecast elements is evaporation, maximum and minimum screen temperature, ground minimum temperature, hours of sunshine, rainfall and probability of rainfall. What guidance is available at any site depends on the site’s observation schedule from 1 July 1996 to 30 April 2001.
3-hourly forecast guidance for 606 stations are also provided for the first two days for screen temperature, MSLP, dew point temperature, relative humidity, visibility, rainfall, total cloud amount, low cloud amount, wind direction and speed. The smaller number of stations is due to the reduced set of sites that provide 3-hourly observations.
Major improvements are evident in the GASP MOF guidance due to a sufficiently long development data set now being available and the removal of potential predictors below 900 hPa. The same low-level predictors in LAPS MOF have also been discarded. This should make MOF (Version 3) more robust with respect to changes in the LAPS low-level physics and do much to remove the biases introduced in MOF (Version 1).
MOF (Version 3) guidance is now provided in both the traditional operational MOF (message) format and in a database friendly format for ingestion to the Forecast Database. The MOF (Version 3) dataflow is presented in Figure 1 – see Appendix A.
The observational archive is generated within a day from the National Climate Centre (NCC) ADAM database, based on the criteria that the sites used are reporting maximum temperature. These reporting sites, screened to ensure that the observation record history was adequate, were retained and others, with discontinuous record histories, were combined with their predecessor site records.
The GASP MOF guidance consists of daily weather forecast elements out to seven days based on the GASP model. The LAPS MOF guidance consists of daily and 3-hourly weather forecast elements out to two days based on the LAPS375 model. The MOF daily forecast elements are a subset (dependent on observation program) of the weather elements in Table 1.
Table 1: MOF Daily Forecast Elements
| Element | Unit | Code |
| Minimum temperature | °C | MINT |
| Maximum temperature | °C | MAXT |
| Rainfall | mm | RAIN |
| Rainfall Probability | percent | PROB |
| Ground minimum temperature | °C | GMIN |
| Sunshine | hr | SUNS |
| Evaporation | mm | EVAP |
The MOF 3-hourly forecast elements are a subset (dependent on observation program) of the weather elements in Table 2.
Table 2: MOF 3-hourly Forecast Elements
| Element | Unit | Code |
| Dry-bulb temperature | °C | TEMP |
| Rainfall since last observation | mm | PRCP |
| Dewpoint temperature | °C | DEWP |
| Wet-bulb temperature | °C | WETB |
| Relative humidity | percent | RELH |
| Wind direction | degrees | WDIR |
| Wind speed | km/hr | WSPD |
| Total cloud amount | eighths | TOTC |
| Low cloud amount | eighths | LOWC |
| Visibility | km | VISB |
| Mean sea level pressure | hPa | MSLP |
Note: The relative humidity is derived from the dry-bulb and dewpoint temperatures.
Disrupted site observation history caused by a shift in location has been handled by compositing observations with preference to the currently operating site where overlapping occurs. The forecast site is the site number and location of the current observing site so that verification and subsequent bias-correction are possible.
A set of 27 NWP grids (see Table 3) is extracted from the NMOC database for each of the model projections (i.e. GASP +24, +48, +72, +96, +120, +144 and +168; or LAPS +00, +12, +24, +36 and +48). A set of 60 NWP grids is extracted for LAPS in MOF (Version 1), and a set of 46 NWP grids is extracted for GASP in MOF (Version 2).
Table 3: NWP Grids
| Field | Level Type | Level 1 | Level 2 |
| Air_temp | surface | 0 | 0 |
| pres | msl | 0 | 0 |
| prcp | surface | 0 | 0 |
| geop_ht | isbr_lvl | 900 850 700 500 | 0 |
| Air_temp | isbr_lvl | 900 850 700 500 | 0 |
| wnd_ucmp | isbr_lvl | 900 850 700 500 | 0 |
| wnd_vcmp | isbr_lvl | 900 850 700 500 | 0 |
| mix_rat | isbr_lvl | 900 850 700 500 | 0 |
| omega | isbr_lvl | 900 850 700 500 | 0 |
The NWP grids are interpolated to the forecast locations described in the MOF station dictionary. These interpolates represent the predictors used in the development of the MOS coefficients. Additional predictors include trigonometric function of day of the year, and those (i.e. thickness, precipitable water, square of the wind speed, moisture transports) derived from the interpolates.
Each forecast element of the MOF guidance is produced by a unique set of coefficients and hence physical inconsistencies and unrealistic guidance can occur. This often can be traced to a small developmental data set due to a limited set of observations.
The MOF guidance have been post-processed to ensure all visibilities, evaporations, probabilities, cloud amounts etc. are within bounds and all rainfall below 0.2mm is set to zero. No internal consistency checks have been implemented.
MOF guidance is now provided in both the operational MOF format (i.e. the traditional message format) and in a database friendly format for ingestion to the Forecast Database.
The daily MOF message contains the daily MOF guidance for each of the stations in the MOF station dictionary, in the traditional message format. Each station's guidance consists of two header lines describing the site number, station name, message identifier, model date, model time and local date; followed by up to seven lines of the daily MOF guidance. Four letter codes (see Table 1) are used in the MOF message to indicate particular forecast weather elements.
For example:
Daily MOF guidance for Adelaide out to seven days generated from the 15/04/2002 00UTC LAPS and 12UTC GASP model runs in the traditional message format.
023090 ADELAIDE AIFS MOF GUIDANCE LAPS 15/04/2002 00UTC DAY /APR 15 /APR 16 /APR 17
MINT 12 16
MAXT 21 24 25
RAIN 0.0 0.0
PROB 14 27
GMIN 6 9
SUNS 8 8 5
EVAP 3.0 3.3
023090 ADELAIDE AIFS MOF GUIDANCE GASP 15/04/2002 12UTC
DAY /APR 16 /APR 17 /APR 18 /APR 19 /APR 20 /APR 21 /APR 22 /APR 23
MINT 9 11 12 12 8 9 8
MAXT 23 26 28 25 21 22 22
RAIN 0.0 0.0 0.0 0.2 0.0 0.0 0.0
PROB 0 0 23 39 23 0 26
GMIN 1 5 6 6 3 7 2
SUNS 12 11 11 7 7 7 9
EVAP 4.1 3.7 3.7 3.6 3.5 3.2 2.7
The 3-hourly MOF message contains the 3-hourly MOF guidance for each of the stations in the MOF station dictionary, in the traditional message format. Each station’s guidance consists of three header lines describing the site number, station name, message identifier, model date, model time, local date and time (AEST); followed by up to eleven lines of the 3-hourly MOF guidance. Four letter codes (see Table 2) are used in the MOF message to indicate particular forecast weather elements.
For example:
3-hourly MOF guidance for Adelaide out to two days generated from the 15/04/2002 00UTC LAPS model runs in the traditional message format.
023090 ADELAIDE AIFS MOF GUIDANCE LAPS 15/04/2002 00UTC
DAY /APR 15 /APR 16 /APR 17
HOUR 12 15 18 21 00 03 06 09 12 15 18 21 00 03 06 09
TEMP 20 21 18 16 14 14 13 17 22 23 20 17 17 17 17 21
PRCP 0.0 0.0 0.0 0.2 0.0 0.0 0.0 0.0 0.2 0.4 0.4 0.0 0.0 0.7 0.0 0.0
DEWP 9 8 9 9 8 8 9 10 10 9 9 9 9 7 9 10
WETB 14 15 14 12 11 11 11 13 15 16 14 13 13 13 13 15
RELH 50 40 60 70 70 70 80 60 50 40 50 60 60 50 60 50
WDIR 320 220 200 120 090 070 060 040 290 250 210 120 090 050 040 030
WSPD 10 15 10 5 5 5 5 5 10 15 10 5 5 10 5 10
TOTC 4 3 1 1 2 2 3 3 3 4 2 2 3 2 4 6
LOWC 2 1 2 1 1 1 2 2 2 1 1 1 1 0 2 2
VISB 35 40 40 35 35 35 35 40 35 35 35 35 35 35 35 35
MSLP 1024 1022 1022 1022 1022 1021 1022 1023 1022 1020 1020 1020 1020 1019 1019 1020
The daily MOF data file contains the daily MOF guidance in the database friendly format. The data file consists of a header line followed by up to eight lines of the daily MOF guidance for each of the stations in the MOF station dictionary.
For example:
Daily MOF guidance for Adelaide out to seven days generated from the 15/04/2002 00UTC LAPS and 12UTC GASP model runs in the database friendly format.
STNNUM, FCST_DATE, MINT, MAXT, RAIN, PROB, GMIN, SUNS, EVAP 023090, 20020415, , 21, , , , 8,
023090, 20020416, 12, 24, 0.0, 14, 6, 8, 3.0
023090, 20020417, 16, 25, 0.0, 27, 9, 5, 3.3
STNNUM, FCST_DATE, MINT, MAXT, RAIN, PROB, GMIN, SUNS, EVAP
023090, 20020416, , 23, , , , 12,
023090, 20020417, 9, 26, 0.0, 0, 1, 11, 4.1
023090, 20020418, 11, 28, 0.0, 0, 5, 11, 3.7
023090, 20020419, 12, 25, 0.0, 23, 6, 7, 3.7
023090, 20020420, 12, 21, 0.2, 39, 6, 7, 3.6
023090, 20020421, 8, 22, 0.0, 23, 3, 7, 3.5
023090, 20020422, 9, 22, 0.0, 0, 7, 9, 3.2
023090, 20020423, 8, , 0.0, 26, 2, , 2.7
The 3-hourly MOF data file contains the 3-hourly MOF guidance in the database friendly format. The text file consists of a header line followed by up to sixteen lines of the 3-hourly MOF guidance for each of the stations in the MOF station dictionary.
For example:
3-hourly MOF guidance for Adelaide out to two days generated from the 15/04/2002 00UTC LAPS model runs in the database friendly format.
STNNUM, FCST_TIME, TEMP, PRCP, DEWP, WETB, RELH, WDIR, WSPD, TOTC, LOWC, VISB, MSLP
023090, 2002041512, 20, 0.0, 9, 14, 50, 320, 10, 4, 2, 35, 1024
023090, 2002041515, 21, 0.0, 8, 15, 40, 220, 15, 3, 1, 40, 1022
023090, 2002041518, 18, 0.0, 9, 14, 60, 200, 10, 1, 2, 40, 1022
023090, 2002041521, 16, 0.2, 9, 12, 70, 120, 5, 1, 1, 35, 1022
023090, 2002041600, 14, 0.0, 8, 11, 70, 090, 5, 2, 1, 35, 1022
023090, 2002041603, 14, 0.0, 8, 11, 70, 070, 5, 2, 1, 35, 1021
023090, 2002041606, 13, 0.0, 9, 11, 80, 060, 5, 3, 2, 35, 1022
023090, 2002041609, 17, 0.0, 10, 13, 60, 040, 5, 3, 2, 40, 1023
023090, 2002041612, 22, 0.2, 10, 15, 50, 290, 10, 3, 2, 35, 1022
023090, 2002041615, 23, 0.4, 9, 16, 40, 250, 15, 4, 1, 35, 1020
023090, 2002041618, 20, 0.4, 9, 14, 50, 210, 10, 2, 1, 35, 1020
023090, 2002041621, 17, 0.0, 9, 13, 60, 120, 5, 2, 1, 35, 1020
023090, 2002041700, 17, 0.0, 9, 13, 60, 090, 5, 3, 1, 35, 1020
023090, 2002041703, 17, 0.7, 7, 13, 50, 050, 10, 2, 0, 35, 1019
023090, 2002041706, 17, 0.0, 9, 13, 60, 040, 5, 4, 2, 35, 1019
023090, 2002041709, 21, 0.0, 10, 15, 50, 030, 10, 6, 2, 35, 1020
The GASP and LAPS MOF guidance are archived to the tape silo mass store under the SAM software system, with file names:
sam/samnmc/mof/laps375/YYYY/_mof_<nwp>_<type>_<date>_<time>.format
and
sam/samnmc/mof/gasp/YYYY/_mof_<nwp>_<type>_<date>_<time>.<format>
where <nwp> = GASP or LAPS,
<type> = day or 3hr,
<date> = YYYYMMDD
<time> = 00 or 12.
<format> = txt or dat
These files can be used to verify the MOF guidance, and are available from 1 November 2001. (early datafiles in this archive were copied from the original AIFS development archive on serva.ho).
The GASP and LAPS MOF guidance will be also be automatically stored in the Forecast Database. These data can also be used to verify the MOF guidance.
The NMOC operational LAPS MOF system is scheduled to run after the LAPS +48 hour grids are stored in the NEONS/ORACLE database. Similarly, the GASP MOF system runs after the GASP database ingestion has completed. The daily and 3-hourly MOF guidance generated from the GASP and LAPS375 (00UTC and 12UTC) model runs are stored in AMDISS and the AIFS Product Store (Regional Offices).
There are 12 products (text and data files, at 00 and 12UTC, for 3-hourly LAPS, daily LAPS, and daily GASP). Each file contains ALL Australia-wide stations.
IDY02110: MOF3 text forecasts (LAPS 3-hourly) 00UTC - Australia
IDY02111: MOF3 text forecasts (LAPS daily) 00UTC - Australia
IDY02112: MOF3 text forecasts (GASP daily to 7 days) 00UTC - Australia
IDY02113: MOF3 text forecasts (LAPS 3-hourly) 12UTC - Australia
IDY02114: MOF3 text forecasts (LAPS daily) 12UTC - Australia
IDY02115: MOF3 text forecasts (GASP daily to 7 days) 12UTC - Australia
IDY02116: MOF3 data forecasts (LAPS 3-hourly) 00UTC - Australia
IDY02117: MOF3 data forecasts (LAPS daily) 00UTC - Australia
IDY02118: MOF3 data forecasts (GASP daily to 7 days) 00UTC - Australia
IDY02119: MOF3 data forecasts (LAPS 3-hourly) 12UTC - Australia
IDY02120: MOF3 data forecasts (LAPS daily) 12UTC - Australia
IDY02121: MOF3 data forecasts (GASP daily to 7 days) 12UTC - Australia
Verification statistics (Tables 4 and 5) have been supplied by Frank Woodcock of BMRC. Verification statistics show that there has been little change in the overall performance for the LAPS MOF output, but the results should be less sensitive to any changes in the model formulation and not show the biases that occasionally occur in the earlier version. The limited results for GASP shown in the Table are better than the current operational system, but only at about the same level as using climatology at the longer forecast ranges.
The verification results are based on a short period and are unstable. Some problems are evident such as the errors for Darwin, which are significantly higher than for climatology. Work in progress indicates that the errors may be significantly reduced with a running bias correction.
Table 4:RMS errors for GASP MOF3 12 UTC capital city forecasts of maximum and minimum temperature for the period 2 December 2001 to 1 January 2002.
| Perth | Darwin | Adelaide | Brisbane | Sydney | Canberra | Melbourne | Hobart | ||
|---|---|---|---|---|---|---|---|---|---|
| Max. | Day 1 | 2.1 | 2.3 | 2.0 | 3.1 | 5.3 | 3.0 | 4.0 | 1.9 |
| Day 2 | 2.5 | 2.9 | 3.1 | 3.2 | 4.7 | 2.7 | 3.2 | 2.5 | |
| Day 3 | 2.4 | 1.4 | 3.6 | 2.4 | 4.1 | 3.3 | 3.4 | 4.2 | |
| Day 4 | 3.0 | 2.6 | 3.3 | 2.3 | 3.2 | 3.4 | 3.6 | 3.5 | |
| Day 5 | 2.9 | 2.6 | 3.6 | 2.1 | 3.7 | 3.3 | 5.3 | 3.1 | |
| Day 6 | 3.9 | 1.5 | 4.1 | 2.3 | 3.9 | 4.2 | 5.1 | 3.0 | |
| Min. | Day 1 | 3.0 | 2.8 | 2.9 | 3.7 | 1.5 | 5.4 | 2.5 | 3.7 |
| Day 2 | 3.9 | 2.2 | 2.6 | 2.9 | 1.4 | 5.6 | 2.1 | 2.9 | |
| Day 3 | 4.3 | 2.8 | 2.7 | 3.1 | 1.8 | 5.6 | 2.5 | 2.7 | |
| Day 4 | 3.5 | 2.2 | 2.0 | 2.5 | 2.0 | 5.0 | 2.2 | 2.9 | |
| Day 5 | 2.5 | 1.5 | 2.1 | 3.1 | 2.5 | 4.0 | 2.1 | 2.5 | |
| Day 6 | 2.6 | 1.8 | 2.5 | 2.5 | 2.2 | 3.7 | 2.2 | 2.4 |
Table 5: RMS errors for LAPS MOF3 00 UTC capital city forecasts of maximum, minimum, ground minimum temperatures, hours of sunshine and evaporation for 60 days until 27 March 2002.
| Perth | Darwin | Adelaide | Brisbane | Sydney | Canberra | Melbourne | Hobart | ||
|---|---|---|---|---|---|---|---|---|---|
| Max. | Day 0 | 2.2 | 1.2 | 1.7 | 1.8 | 1.6 | 2.3 | 2.4 | 2.6 |
| MOF3
Old |
Day 1 | 2.0
2.4 |
3.9
1.0 |
2.0
1.9 |
1.4
1.7 |
2.2
2.1 |
2.8
4.8 |
2.5
3.9 |
2.6
2.9 |
| Day 2 | 2.8 | 1.3 | 3.2 | 1.3 | 2.4 | 2.7 | 3.4 | 2.3 | |
| Min.
Old |
Day 1 | 3.7
1.9 |
1.7
1.6 |
2.5
3.5 |
1.1
2.0 |
1.4
2.1 |
2.6
4.6 |
2.4
4.1 |
2.6
2.3 |
| Day 2 | 3.1 | 1.6 | 2.6 | 1.4 | 1.5 | 2.8 | 3.1 | 2.6 | |
| Sun | Day 0 | 2.0 | 2.2 | 2.3 | 2.8 | 3.5 | 2.4 | 2.5 | 3.1 |
| Day 1 | 2.4 | 2.6 | 2.6 | 3.5 | 3.6 | 2.4 | 2.9 | 2.7 | |
| Day 2 | 2.6 | 2.8 | 3.1 | 3.6 | 3.5 | 2.6 | 2.8 | 2.9 | |
| Evap | Day 1 | 1.6 | 2.4 | 1.8 | 2.3 | 2.5 | 1.9 | 3.3 | 2.2 |
| Day 2 | 1.7 | 3.1 | 1.6 | 2.5 | 2.4 | 1.8 | 8.2 | 2.2 | |
| Gmin | Day 1 | 3.7 | 1.3 | 3.0 | 2.2 | 2.4 | 3.0 | 3.9 | 4.4 |
| Day 2 | 3.2 | 1.6 | 3.5 | 2.5 | 2.2 | 3.2 | 3.8 | 4.1 |
Note: Results for Day+1 maxima and minima forecasts from the new MOF3 and the current MOF are included in Table 5. The current MOF forecasts are from the last 30 days AIFS FVS results while the MOF 3 results are from the last 60 days.
The MOF (Version 3) system dataflow is presented in Figure 1.
Figure 1: MOF (Version 3) System Dataflow
The MOF (Version 3) forecast locations are presented in Figure 2.
Figure 2: MOF (Version 3) Forecast Locations
Home | About Us | Learn about Meteorology | Contacts | Search | Help | Feedback Weather and Warnings | Climate | Hydrology | Numerical Prediction | About Services | Registered Users |
|
© Copyright Commonwealth of Australia 2009, Bureau of Meteorology (ABN 92 637 533 532) Please note the Copyright Notice and Disclaimer statements relating to the use of the information on this site and our site Privacy and Accessibility statements. Users of these web pages are deemed to have read and accepted the conditions described in the Copyright, Disclaimer, and Privacy statements. Please also note the Acknowledgement notice relating to the use of information on this site. No unsolicited commercial email. |
