Climatic Research Unit: Data

Lamb Weather Types (LWTs)

Original LWTs

Reanalysis based LWTs - UK and N. Sea

Gale index from reanalysis based LWTs - UK and N. Sea

Flood Indices based on LWTs


Professor H.H. Lamb subjectively classified each day's weather over the British Isles from 1861 to February 1997. All the years have been published, except for 1996/7, in Lamb (1972) and Hulme and Barrow (1997). The complete series of the classification is given here. Several periods before 1861 have also been classified (see Lamb, 1991).

File format
format(a) descriptive header line
for year = firstyear to lastyear (rows)
for month = 1 to 12 (columns)
format(31i3) each day of month

Lamb Types - Number coding
-1  U           -9  non-existent day
0 A 20 C
1 ANE 11 NE 21 CNE
2 AE 12 E 22 CE
3 ASE 13 SE 23 CSE
4 AS 14 S 24 CS
5 ASW 15 SW 25 CSW
6 AW 16 W 26 CW
7 ANW 17 NW 27 CNW
8 AN 18 N 28 CN

Filename Description
lwtlamb.dat Lamb Weather Types - Lamb 1861-1997



An objective scheme to classify the daily circulation according to the Lamb weather typing scheme was developed by Jenkinson and Collison (1977). The objective scheme uses daily grid-point mean sea level pressure data (see map below). The objective and the original subjective Lamb scheme have been compared by Jones et al., (1993). A new series has now been producuced using reanalyses data (Jones, et al., 2013). For the period from 1871-1947 the 20CR developed by Compo et al. (2011) is used and for 1948 to present the National Centers for Environmental Prediction, NCEP Reanalysis by Kalnay et al., (1996) is used. In developing a new and more consistent series, we have chosen to base the analysis data on the synoptic hour 12, as this is in the centre of the civil day to which the LWTs refer. The series from 1871 to the present day is given here.

File format
Description of columns
day month year PM-1000 W S F Z G Dir LWT
PM_1000 average pressure over the grid points
W westerly flow
S southerly flow
F resultant flow
Z total shear vorticity
G gale day
Dir direction of flow
LWT see above table

Grid centred on the UK

Location of the grid points over the British Isles.

Filename Description
20CR_1871-1947_ncep_1948-2017_12hrs_UK.dat 1871-1947 20CR; 1948-present NCEP
20CR_1871-1947_ncep_1948-2017_12hrs_UK.csv as above in csv format (contains headers)
20CR_1948-2010_12hrs_UK.dat 1948-2010 (20CR to facilitate comparison)

The following files provide the output described at the other synoptic hours, but they do not contain the LWT:

Filename Description
ncep_1948_2017_00hrs_UK.dat 1948-present NCEP 00hrs
ncep_1948_2017_06hrs_UK.dat 1948-present NCEP 06hrs
ncep_1948_2017_18hrs_UK.dat 1948-present NCEP 18hrs

Filename Description
20CR_1871-2010_00hrs_UK.dat 1871-2010 20CR 00hrs
20CR_1871-2010_06hrs_UK.dat 1871-2010 20CR 06hrs
20CR_1871-2010_18hrs_UK.dat 1871-2010 20CR 18hrs

UK Jenkinson Gale Index

The gale index G found in the above files is calculated from the following formula:

                G = [F² + (0.5Z)²]1/2

where F is the resultant flow and Z is the total shear vorticity, see Jones et al (1993) for further details. In a similar vein to Hulme and Jones (1991) exceedances of a threshold have been counted and allocated a gale intensity:

G > 30 gale

G > 40 severe gale

G > 50 very severe gale

This was done annually and for an extended winter (NDJFMA)

Filename Description
gale_index_1200hrs_UK.csv dated log of all G>30 events
ANN_G_thresh_counts_UK.pdf Annual plots of the three thresholds
NDJFMA_G_thresh_counts_UK.pdf Extended winter plots of the three thresholds

Grid centred on the N. Sea

Location of the grid points over the North Sea.

Filename Description
20CR_1871-1947_ncep_1948-2017_12hrs_N_Sea.dat 1871-1947 20CR; 1948-present NCEP
20CR_1871-1947_ncep_1948-2017_12hrs_N_Sea.csv as above in csv format (contains headers)
20CR_1948-2010_12hrs_N_Sea.dat 1948-2010 (20CR to facilitate comparison)

The following files provide the output described at the other synoptic hours, but they do not contain the LWT:

Filename Description
ncep_1948_2017_00hrs_N_Sea.dat 1948-present NCEP 00hrs
ncep_1948_2017_06hrs_N_Sea.dat 1948-present NCEP 06hrs
ncep_1948_2017_18hrs_N_Sea.dat 1948-present NCEP 18hrs

Filename Description
20CR_1871-2010_00hrs_N_Sea.dat 1871-2010 20CR 00hrs
20CR_1871-2010_06hrs_N_Sea.dat 1871-2010 20CR 06hrs
20CR_1871-2010_18hrs_N_Sea.dat 1871-2010 20CR 18hrs

N. Sea Jenkinson Gale Index

See UK section for explanation.

Filename Description
gale_index_1200hrs_N_Sea.csv dated log of all G>30 events
ANN_G_thresh_counts_N_Sea.pdf Annual plots of the three thresholds
NDJFMA_G_thresh_counts_N_Sea.pdf Extended winter plots of the three thresholds


Fluvial flood indices for British regions and the Island of Ireland (1871-present)

The file mov_averages.pdf contains flood indices based on the method of Wilby and Quinn (2013) for Environment Agency regions in England and Wales, plus the area overseen by the Scottish Environmental Protection Agency (SEPA). The table below contains the data used to produce the plots including the complete time series from which the moving averages were calculated.

Filename Description
day_weights.csv Complete time series
move_av_30dy.csv 30 day moving average
move_av_90dy.csv 90 day moving average
move_av_365dy.csv 365 day moving average

The map below shows the regions (by colour) and river gauging stations used to calibrate the flood indices. Stations were selected because they had at least 50 years of flood peak data.

Gauging stations map

Flood indices are reconstructed from the number of Peaks over Threshold (POT) river flows in relation to the prevailing daily Lamb Weather Types (LWTs). POT time series are derived for each station by determining how many times the river flow exceeds a threshold set for each gauge, recorded as the number of exceedances per water year (October to September).

The LWTs used by Wilby and Quinn (2013) are based on an objective classification of pressure patterns obtained from the 20CR and NCEP/NCAR re-analysis described above. Like the regional flood indices, these LWTs are provided from 1871 to present, updated daily.

Weather patterns such as the C-type tend to generate more frequent POTs whereas A-types are relatively flood poor. Hence, decadal variations in the frequency of these contrasting weather types can explain flood rich and flood poor periods. Since gauged flood data are rare (very few stations were operating prior to 1950), the association between recorded POTs and corresponding LWT enables consideration of flood clustering in time periods where only LWT data are available (1871 to 1950s).

The regional flood indices also help place contemporary events in a much longer context. For example, the 30 day moving average flood index in February 2014 reached a maximum weight value of 2.33 in the Southern region. This value has only been exceeded during five previous episodes in the last 144 years (1872, 1951, 1981, 1983 and 2000).

Overall, the most extreme flood index for the Southern region was 2.58 in November 2000. As with 2014 this high value was due to very persistent cyclonic weather lasting several weeks.

A flood index for the Island of Ireland (IoI) was subsequently produced by Harrigan (2016) following the approach of Wilby and Quinn (2013). The IoI flood index was calibrated over water years 1978-2009 using POTs based on Daily Mean Flows (DMFs) from the 29 stations shown in the map below.

IoI station map

Similar to Britain, the C-type weather pattern generates more frequent POTs across IoI. In general, IoI flood loadings are most similar to those for Wales where the CS-type is also found to be associated with a greater likelihood of high flows.


Last updated: Jan 2017, Colin Harpham

These datasets are made available under the Open Database License.
Any rights in individual contents of the datasets are licensed under the
Database Contents License under the conditions of Attribution and Share-Alike.
Please use the attribution Climatic Research Unit, University of East Anglia