Rainfall over the outcrop areas of most major aquifers during 2003 was 15-20% below the 1961-90 average and very unevenly distributed through the year; typically, around 45% of the annual total fell in January, November and December. Thus, the most abundant rainfall occurred during that part of the year when groundwater replenishment is normally at a maximum. This, combined with the very healthy state of groundwater resources following exceptional replenishment late in 2002, meant that the drought conditions which obtained throughout much of 2003 year had only a limited impact on groundwater resources. A similar annual rainfall total across the major aquifer outcrop areas but with the bulk of the rainfall during the summer half-year would have had a much more damaging impact on groundwater resources – and on the streams supported by the outflows from springs and seepages.
Throughout much of the Chalk, the majority of the limestone aquifers, and the most responsive of the Permo-Triassic sandstones outcrops, the 2002 seasonal recovery in groundwater levels began from a lower base than in any of the previous five years – although only in a few areas (e.g. parts of the southern Chalk) did it approach drought minima. This, together with the delay in the seasonal onset of infiltration in the autumn (a consequence of the dry late summer soil conditions in 2002) suggested that groundwater levels might be appreciably below average at the beginning of 2003. In the event, exceptional rainfall over a period spanning the second week of October to mid-November, rapidly reduced soil moisture deficits and initiated a period of very heavy infiltration which continued until early January 2003. As a consequence, groundwater levels throughout most major aquifers were notably high – and still increasing – at the start of 2003 (see Figure 1).
The behaviour of groundwater levels throughout the year is shown in Figure 2 which features 1999-2003 hydrographs for a selection of index wells and boreholes throughout the UK. Five-year plots have been used because groundwater levels in many areas show considerable persistence – reflecting groundwater replenishment over a number of recharge seasons. The groundwater level trace is shown together with the monthly maximum and minimum levels for the pre-1999 record. For 2003, the hydrographs suggest a near-normal seasonal variation in levels for a few index wells and boreholes. Most however are characterized by unusually protracted recessions and a very large range of levels over the year – exceeding 40 metres in the case of the Chilgrove well in the South Downs. The nine- or ten-month recessions testify to the existence of drought conditions but, a few index sites aside, the annual minimum levels in 2003 were unexceptional. This reflects the limited vulnerability of most aquifer units to drought episodes that are confined to a single year. Although the groundwater resources outlook was a matter of some concern in September and October (see Figure 3), the heavy late autumn and early winter rainfall in 2003 initiated a recovery which continued well into the following year. By late January 2004, overall groundwater resources were well above average.
The majority of observation wells and boreholes for which data are held on the National Groundwater Level Archive monitor the natural variation in levels. However, in parts of the UK groundwater levels have been influenced, sometimes over very long periods, by pumping for water supply or other purposes. As a consequence, some local or regional water-tables have become substantially depressed. For instance, contemporary levels at a number of boreholes in the Permo-Triassic sandstones of the Midlands are indicative of a significant regional decline. In London, increasing groundwater abstraction through the nineteenth and the first half of the twentieth centuries led to a 70-metre decline in groundwater levels in the Trafalgar Square borehole. Since the 1950s, a much reduced abstraction rate has resulted in a recovery of around 30 metres with levels rising by 1-2 metres a year through the early 1990s (Figure 4). Level data for the 1998-2002 period are under review but the 2003 data confirm that control measures (principally abstraction by Thames Water) instigated in the late 1990s have reversed the increase – other wells in central London confirm a decline in levels since around 1999. Rising groundwater levels have also been reported from other conurbations; leakage from water mains is considered a significant factor in some cases. The implications of rising groundwater levels extend beyond the potential improvement in water resources that the rise represents. Groundwater quality may be adversely affected as levels approach the surface and a number of geotechnical problems may result, for instance the flooding of tunnels and foundations.
Groundwater levels approached seasonal maxima throughout many major aquifer units in January 2003. Echoing conditions in early 2001, groundwater flooding was experienced in some Chalk outcrop areas. With catchments saturated and rainfall broadly favouring the eastern outcrop areas (for the seventh successive month in some localities), January infiltration totals exceeded the average throughout most major aquifers. Infiltration over the four months beginning in September 2002 was more notable, exceeding twice the average throughout a significant proportion of the eastern Chalk (e.g. in parts of the North Downs and Chilterns). Only in 2000/01 has there been an appreciably more abundant autumn/early winter groundwater replenishment over the post-1960 period. The impact of these two exceptional recharge episodes is evident in the groundwater level hydrographs (Figure 1). January 2003 levels in most Chalk outcrop areas were unprecedented in a few areas (e.g. parts of the Yorkshire Wolds and at Rockley). Levels were notably high in most of the (more responsive) limestone outcrops and still exceeded pre-2000 maxima in a few parts of the slower-responding Permo-Triassic sandstones (e.g. at Llanfair DC).
February rainfall totals were below 50% of the monthly average across large parts of the Chalk outcrop and a dry, mild spell in mid-month saw modest soil moisture deficits established in some eastern and southern areas. Infiltration rates for February were typically only 25-50% of average, helping to avoid a repeat of the protracted groundwater flooding which occurred in early 2001. Significant declines in groundwater levels were reported for a number of the more responsive Chalk wells and boreholes (e.g. at Chilgrove and Rockley); nonetheless, levels generally remained notably high (Killyglen in Northern Ireland was an exception). In many slower-responding wells (e.g. Therfield), levels were still rising and exceeded pre-2000 February maxima in some areas, including the Chilterns; correspondingly spring outflows were notably heavy. Groundwater levels in the Middle Jurassic limestone of the Cotswolds fell to within the normal late-winter range but in most limestone outcrops, late-winter levels were close to seasonal maxima. Spatial variability was considerable across the Permo-Triassic sandstone outcrops but, aside from wells affected by abstraction, levels remained high – exceptionally so for some index wells in North Wales, the North-West and the Midlands.
Infiltration in March was again modest – less than 30% of the monthly average in some eastern aquifer outcrop areas. However, as is often the case in the spring, the differing responsiveness of individual aquifer units presented (seemingly) conflicting signals. In the Chalk, steep late winter/early spring declines in groundwater levels at some index boreholes (e.g. West Woodyates and Rockley) contrasted with continuing recoveries in some of the deeper eastern wells. Despite the paucity of late-winter/early-spring recharge, groundwater levels remained above average throughout most of the Chalk aquifer. Brisk recessions in most limestone aquifers left late-spring groundwater levels close to the seasonal average. As a consequence of the high storage capacity in many outcrop areas, levels in many Permo-Triassic sandstones outcrops reflect recharge patterns over several years. Consequently, March levels remained very high in most northern outcrops and in the Midlands. Overall groundwater resources for England and Wales also remained healthy at month end.
In April, a combination of low rainfall, high evaporative demand and correspondingly high soil moisture deficits terminated the recharge season in many aquifer outcrop areas – particularly those in eastern England. The spring groundwater level recessions continued, but varied greatly in their steepness. The slower responding Chalk wells (e.g. in parts of the Chilterns) were still exhibiting water levels close to their 2002/03 maxima, whilst in the far west of the Chalk outcrop some wells showed appreciably more depressed levels. The Chalk in Northern Ireland showed the greatest apparent diminution in storage – partly a consequence of the shortness of the Killyglen record (19 yrs). The Permo-Triassic sandstones also did not present a spatially coherent picture. Wells in the West Midlands and Lancashire reported levels well above average – some were close to the maximum recorded – whereas in the Eden valley (Cumbria), Devon, North Wales and Nottinghamshire, levels were within the normal April range. The status of groundwater resources in late April was generally satisfactory.
Seasonally very high soil moisture deficits throughout May confirmed the end of the 2002-03 recharge season across almost all outcrop areas of the major aquifers. Infiltration during the month was minimal (thunderstorms produced some very localised infiltration) and the drought conditions throughout much of the spring resulted in March-May recharge totals of less than a quarter of the long term average across much of the eastern Chalk. May groundwater levels contrasted sharply with those at the beginning of the year but, although groundwater levels were relatively depressed in the south-western outcrops, levels remained in the normal range across most of the Chalk aquifer. A similar picture characterised levels in a number of minor aquifers (e.g. the Norfolk Drift and Essex Gravels). The 2003 spring recession was particularly steep in the limestone aquifers – at Ampney Crucis in the Jurassic limestone of the Cotswolds only during the 1976 and 1984 droughts have lower May levels been recorded. Steep spring declines were evident in the more responsive Permo-Triassic sandstones units but, generally, the large storage within the aquifer – which militates against rapid change – meant that levels were still well above the early summer average.
With evaporative demands again well above average over wide areas soils in most most aquifer outcrop areas were exceptionally dry by late June. Infiltration during June was thus largely confined to very localised episodes associated with notable storm rainfall totals (e.g. a modest increase in groundwater levels was reported for the Carboniferous Limestone at Alstonfield in the north Midlands). Generally however the spring/summer recessions continued – by month-end levels in parts of the western Chalk (e.g. at Woodyates and Chilgrove) had fallen around 30m from their January maxima, and were appreciably below the average for early July. But levels in the majority of index wells in the Chalk were still well within the normal early summer range. Although a similar generalisation applied to most limestone outcrops, however, late-June levels at Ampney Crucis in the Cotswolds were the lowest since June 1997. Following more than two years above previous seasonal maxima in some areas (e.g. North Wales), the 2003 recession returned levels in most of the Permo-Triassic sandstones aquifer to the normal range, albeit still well above average in the northerly outcrops. Levels in the minor aquifers of eastern England continued to follow a typical summer recession.
Notwithstanding the late July rainfall, soil moisture deficits were considerably above average across most major outcrop areas by month-end, and increased further in early August. Infiltration during July was, as usual, minimal and groundwater level recessions continued in all the major aquifers. Recessions in the Chalk generally followed a typical shallow summer decline but groundwater levels in the most southerly outcrops were depressed. Levels were also well below the average in the Jurassic Limestone of the Cotswolds but generally within the normal range in the other limestone aquifers. In the Permo-Triassic sandstones, levels in the Llanfair DC and Heathlanes index boreholes closely approached the average for first time since 1999; the Yew Tree Farm hydrograph also exhibited a sustained decline but had only recently fallen below pre-2000 maxima. Overall groundwater resources for England and Wales were around the average for the time of year.
As is normally the case, many of the highest late summer soil moisture deficits coincided with the outcrop areas of the major aquifers. Correspondingly, the summer groundwater level recessions continued – leaving levels particularly depressed in parts of the southern Chalk. At Compton (in the South Downs), where levels have been measured since 1894, lower August levels have only been recorded during the severe drought of 1975/76. Chalk levels were also low in Northern Ireland but throughout the greater part of the outcrop they remained within the normal late summer range. Levels in the limestone aquifers similarly remained within the normal range but considerably below average whilst levels remained healthy throughout almost all of the Permo-Triassic sandstones outcrops. At month-end, soil moisture deficits were the equivalent of more than12 weeks residual rainfall (rainfall-actual evaporation) in some eastern areas – further declines in groundwater levels were therefore inevitable.
With record soil moisture deficits across many outcrop areas, infiltration during September was again minimal and the, already protracted, 2003 groundwater level recessions continued. Groundwater levels were especially depressed in parts of the southern Chalk whilst, to the east and north, levels in the Chalk remained considerably healthier – still above average in parts of the Chilterns and mostly within the normal early autumn range elsewhere. Generally, the limestone aquifers presented a similar picture, as did a number of the minor eastern aquifers (e.g. the Essex gravels and Norfolk Drift). Levels in most of the southern Permo-Triassic sandstones outcrops also remained within the normal range but the slower-responding Midland and northern outcrops were still close to pre-2000 maxima, albeit at their lowest for four or five years in most cases.
Rainfall over most outcrop areas was above average in October and soil moisture deficits decreased briskly around month end. However, deficits were still amongst the highest on record (for October) across most aquifer outcrop areas; consequently recharge was minimal. The absence of significant recharge over the Feb-Oct period had produced a dramatic decline in groundwater levels – one with few modern parallels (1995 was comparable). But there were important spatial variations in the health of groundwater resources. In the south-western Chalk, water-tables in some areas were approaching natural base levels (e.g. at Chilgrove) and a number of wells had been reported as dry. To the east and north, groundwater levels in the Chalk were mostly below average but still within the normal range; a few (e.g. in the Chilterns) remained a little above average. The delayed onset of the 2003 seasonal recovery was particularly evident in the limestone aquifers where October levels were generally well below average – but still considerably above drought minima. This was true of many minor aquifers also, and most boreholes in the slower responding Permo-Triassic sandstones outcrops reported levels within the normal late-autumn range. However, concern focused on the possibility that the 2003/04 recharge season could be further truncated by limited rainfall over the late autumn.
Late-October and early-November rainfall rapidly reduced outstanding soil moisture deficits allowing infiltration to recommence in some of those areas where the water-tables were most depressed (the southern Chalk especially). But soils remained seasonally very dry across parts of the Midlands, Yorkshire and East Anglia (coastal areas of north-east Britain also). Reporting dates for many index boreholes meant that the full effect of the late-November infiltration was not evident on the November groundwater level traces. Nonetheless, groundwater level increases were recorded (from a very low base) at index wells in the Jurassic Limestone, Magnesian Limestone and in the south-western Chalk. Recessions continued across much of the rest of the Chalk outcrop – at Chilgrove, late-November levels were the 3rd lowest in a continuous series from 1836. The degree of water-table depression was more modest to the north but, generally, levels in the Chalk were below any recorded since 1996 or 1997. This was also true of the more responsive Permo-Triassic sandstones outcrops but levels remained very healthy across most of the outcrops – although spatial variations were large reflecting differing recharge patterns and aquifer characteristics. Overall groundwater resources were considerably below average for the autumn but significantly healthier than at the same time in 1995 and 1996.
The frontal rainfall, mostly of modest intensity, during December helped to consolidate the groundwater recovery that had gathered momentum in the late autumn. A significant proportion of index wells reported too early in December to fully capture the consequent groundwater level recovery but notable rises were reported for a number of responsive aquifer units (particularly in the limestone aquifers, but in the Isle of Wight Chalk also). The upturn in the southern Chalk was especially welcome and levels in most index wells and boreholes returned to within the normal range – albeit still low for the early winter. Even in the normally late-responding eastern aquifer units (e.g. the Essex Chalk), and in the minor aquifers (e.g. the Suffolk Crag) modest rises were recorded in December. In the context of the abundant 1998-2002 recharge, the paucity of 2003 infiltration to the slowest responding Permo-Triassic sandstones outcrops is of limited significance – levels in most index wells remained above average at year end. Near-saturated soils in almost all outcrop areas ensured that further significant infiltration would occur in early 2004.