Home
Up
Water Resources
Rainfall Modelling
Climate variability
Historical Drought

Climate Variability

 

Published in:

 

Fowler, H.J. and Kilsby, C.G. 2002. Precipitation and the North Atlantic Oscillation: A study of climatic variability in Northern England. Int. J. Climatol., 22, 843-866. [Abstract]

 

The North Atlantic Oscillation (NAO) index is a measure of pressure between Iceland and the Azores and the usual index is an average of the December to March values. Here, we have used the NAO index calculated at the Climatic Research Unit, University of East Anglia, UK (Figure1). We know that shifts in the NAO have a profound effect on temperature and precipitation in northern Europe by way of changes in synoptic weather patterns. These are particularly associated with changes in the frequency and magnitude of surface westerlies across northern Europe. There has been a  recent upward trend in the NAO, in fact from the 1960s to the early 1990s, but there is also high year-to-year variability.

Figure 1 The North Atlantic Oscillation index (NAO) for winter (DJFM). Imposed upon the yearly values is an 11-yr centred moving average that shows the recent increase in the index.

 

 

So how does this affect climate in Yorkshire (and elsewhere in the UK?)

 

Well, our research has shown that climate has changed, particularly over the past 30-40 years. In winter, we have seen an increase in precipitation since the late 1970s (Figure 2) which can be linked to recent upward trend in the NAO index. In summer, however, we see a contrasting trend with a significant decrease in precipitation since the 1960s (Figure 3).

Figure 2 Winter anomaly  in precipitation at Moorland Cottage (western Yorkshire, UK), expressed as percent anomalies from the 1937-1996 average

Figure 3 Summer anomaly  in precipitation at Moorland Cottage (western Yorkshire, UK), expressed as percent anomalies from the 1937-1996 average

 

And, in fact, this variability is significantly linked to atmospheric circulation patterns such as the NAO. At western sites, precipitation is highly correlated to the NAO, with positive and negative winter-NAO indices producing anomalies from the long-term winter precipitation average of 112 and 77 percent respectively. At eastern sites, less significant relationships are found (see Table 1)

 

Table 1 (a) Mean rainfall amounts; (b) Mean proportion dry days; both annually and seasonally, associated with positive and negative NAO. Numbers in parentheses indicate the percentage of the long-term average rainfall. Numbers both bold and italicised are from significantly different populations when positive and negative NAO values are compared (95 percent confidence interval).

 

(a)

Site

Ann-P

DJF-P

MAM-P

JJA-P

SON-P

 

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

Hury

Reservoir

942

(102)

 

907

(98)

289

(107)

246

(91)

181

(96)

206

(109)

203

(100)

201

(100)

278

(104)

247

(93)

Barnard

Castle

786

(100)

 

775

(99)

229

(104)

203

(93)

153

(94)

181

(111)

179

(99)

183

(101)

229

(105)

199

(91)

Raby

Castle

746

(100)

 

744

(100)

198

(100)

200

(101)

150

(94)

176

(111)

188

(100)

186

(99)

217

(105)

185

(90)

Lockwood

Reservoir

831

(100)

 

842

(101)

191

(97)

211

(107)

167

(98)

177

(104)

221

(97)

242

(107)

248

(103)

222

(92)

Scarborough

655

(100)

 

650

(99)

163

(97)

179

(106)

131

(97)

145

(107)

171

(101)

168

(99)

186

(104)

164

(92)

Moorland

Cottage

1959

(104)

 

1722

(92)

636

(112)

442

(77)

360

(99)

370

(102)

394

(100)

377

(96)

567

(102)

505

(91)

Kirk

Bramwith

585

(99)

 

598

(102)

136

(100)

134

(99)

125

(99)

132

(105)

167

(97)

182

(106)

155

(101)

152

(99)

(b)

Site

Ann-PD

DJF-PD

MAM-PD

JJA-PD

SON-PD

 

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

+NAO

-NAO

Hury

Reservoir

0.4

(101)

 

0.4

(97)

0.3

(97)

0.4

(103)

0.5

(104)

0.4

(91)

0.5

(101)

0.5

(98)

0.4

(99)

0.4

(102)

Barnard

Castle

0.5

(100)

 

0.5

(100)

0.4

(98)

0.4

(104)

0.5

(103)

0.5

(94)

0.5

(101)

0.5

(99)

0.4

(97)

0.5

(105)

Raby

Castle

0.5

(101)

 

0.5

(99)

0.4

(98)

0.5

(104)

0.6

(104)

0.5

(93)

0.6

(101)

0.6

(99)

0.5

(98)

0.5

(104)

Lockwood

Reservoir

0.5

(101)

 

0.4

(98)

0.4

(101)

0.4

(97)

0.5

(101)

0.5

(97)

0.5

(102)

0.5

(96)

0.4

(99)

0.4

(102)

Scarborough

0.5

(100)

 

0.5

(99)

0.4

(101)

0.4

(97)

0.6

(102)

0.5

(96)

0.6

(101)

0.6

(98)

0.5

(98)

0.5

(105)

Moorland

Cottage

0.4

(100)

 

0.4

(100)

0.4

(94)

0.5

(113)

0.5

(102)

0.4

(95)

0.5

(98)

0.5

(98)

0.4

(98)

0.4

(99)

Kirk

Bramwith

0.6

(101)

 

0.5

(98)

0.5

(99)

0.5

(100)

0.6

(101)

0.6

(98)

0.6

(102)

0.6

(96)

0.5

(101)

0.5

(100)

 

These 'shifts' in rainfall during a positive or negative NAO phase can be linked to changes in the frequency of rain-bearing weather patterns. In Yorkshire, weather can be categorised into three distinct weather 'states': westerly, easterly and anticyclonic. In meteorological terms these may roughly be defined as zonal, meridional, and blocking respectively. Westerly weather provides the west of the region with most of its rainfall, whereas the east receives its rainfall mainly from easterly weather. Anticyclonic conditions provide dry weather across the region.

 

These can be further split into seasonal types: winter and summer. This gives 6 distinct weather patterns and clear correlation can be found between these weather states and the NAO.During a high winter-NAO period, such as from 1980-1990, there is an increased frequency of the WW and SA weather states, to the detriment of the SW and WA weather states. In a low winter-NAO period, such as 1960-1970, the reverse situation occurs.

 

A high correlation is obtained between monthly NAO and precipitation for winter months (see Figure 4a). A positive relationship is also found between monthly NAO and the occurrence of the westerly weather state in all months. The correlation is particularly strong in the month of March, at 0.932 (see Figure 4b). In March, the westerly weather state has increased in frequency from 16 to 24 days on average since 1975: a possible consequence of the positive NAO index since that time.

 

 

 

Figure 4 (a) Monthly NAO and precipitation, and (b) frequency of occurrence of a weather state, for Moorland Cottage in March, using data from 1937-1998 (11-yr centred moving average).

 

This research has shown that, in Yorkshire, the NAO, the frequency of weather types and precipitation totals are intimately linked on an annual, seasonal and monthly time-scale. Strong connections have been found between winter precipitation totals and the winter-NAO, particularly in western Yorkshire where most surface water supplies are located. These may be very important in the future prediction of winter rainfall. If the NAO can be forecast with some skill using sea surface temperatures or other variables then the capability may exist to estimate the winter replenishment of many Pennine reservoirs a few months in advance.

It is likely that similar relationships exist in other parts of the western UK, for example Ireland and north-west England. An extension of this research to include such areas would provide an important predictor of future hydrological drought episodes and allow the forward planning and management of many water resource systems located in climatologically similar regions of the UK.