The influence of recent environmental change ... - Wiley Online Library

0 downloads 0 Views 332KB Size Report
Yorkshire Ouse basin. Sean A. Longfield1* and Mark G. Macklin2. 1School of Geography, University of Leeds, Leeds, LS2 9JT, UK. 2Institute of Geography and ...
HYDROLOGICAL PROCESSES Hydrol. Process. 13, 1051±1066 (1999)

The in¯uence of recent environmental change on ¯ooding and sediment ¯uxes in the Yorkshire Ouse basin Sean A. Longfield1* and Mark G. Macklin2

2Institute

1School of Geography, University of Leeds, Leeds, LS2 9JT, UK of Geography and Earth Sciences, University of Wales Aberystwyth, Aberystwyth, SY23, Wales, UK

Abstract: This study attempts to establish causes of historical variations in ¯ood frequency and magnitude in a 119-year ¯ood stage record at York, and assess its likely e€ects on ¯uxes of ®ne sediment and associated base metal contaminants. Climatic controls of ¯ooding are evaluated through analysis of synoptic ¯ood generation, and large-scale land use changes are also considered in terms of their in¯uence on food frequency and magnitude and sediment availability. The late nineteenth century was characterized by low ¯ood frequency and magnitude, but, as a consequence of upland metal mining, contaminant concentrations and downstream ®ne sediment delivery would have been high. A further decline in ¯ood frequency and magnitude between 1904 and 1943, combined with the cessation of base metal mining, resulted in a decline of sediment and contaminant ¯uxes. Between 1944 and 1968 increases in upland drainage, changes in agricultural practices and high ¯ood frequency and magnitude resulted in signi®cantly enhanced sediment ¯uxes. Rates of ¯ux declined between 1969 and 1977 owing to extremely low ¯ood frequencies and magnitudes. In the last two decades there have been a series of extreme magnitude ¯oods, which through remobilization of mining-contaminated alluvium have resulted in high pollutant metal mining loads, probably approaching those of the late nineteenth century. In an historical context it appears that the LOIS monitoring programme has been undertaken in a period of extremely high ¯ood frequency and magnitude in the UK, and signi®cantly enhanced sediment and contaminant metal ¯ux. Analysis of synoptic ¯ood generation suggests that any future increase in the frequency and vigour of cyclonic atmospheric circulations may result in a higher frequency of extreme ¯oods and consequently increased sediment ¯uxes. Copyright # 1999 John Wiley & Sons, Ltd. KEY WORDS

LOIS; Ouse at York; ¯ooding; climate change; atmospheric circulation; synoptic ¯ood generation; land use change; sediment ¯uxes; environmental change

INTRODUCTION Sediment ¯uxes and transport of associated particulate-borne metal contaminants (e.g. Cd, Pb, Zn) signi®cantly increase during ¯ood events (Fox and Johnson, 1997; Macklin et al., 1997; Wass et al., 1997; Webb et al., 1997). The largest ¯ood during the LOIS monitoring programme (1993±1996) occurred in winter 1995 with new peak discharges being registered on the rivers Ure and Wharfe (NERC, 1996; Law et al., 1997; Marsh and Sanderson, 1997), both with records in excess of 30 years. The main aim of this paper is to establish longer term variations in ¯ood frequency and magnitude and assess the likely e€ects on ¯uxes of ®ne sediment and associated contaminants. Climatic controls of ¯ooding are evaluated by determining the * Correspondence to: Dr. S. A. Long®eld, School of Geography, University of Leeds, Leeds LS2 9JT, UK. E-mail: [email protected] Contract grant sponsor: UK NERC. CCC 0085±6087/99/071051±16$1750 Copyright # 1999 John Wiley & Sons, Ltd.

Received 1 January 1998 Revised 21 April 1998 Accepted 21 April 1998

1052

S. A. LONGFIELD AND M. G. MACKLIN

synoptic weather conditions that generate ¯oods, and how these vary for di€erent magnitude events. Large-scale land use changes, including land draining, changing agricultural practices and base metal mining, are also considered in terms of their e€ects on runo€ and sediment availability in the last 120 years. STUDY AREA The River Ouse at York (Figure 1) is non-tidal and has a catchment area of approximately 3315 km2. It has three major tributaries, the Swale, Ure and Nidd, which ¯ow in a south-easterly direction from the Yorkshire Dales National Park. The area upstream of York is predominantly rural agricultural land with a low population density (Jarvie et al., 1997). Altitude in the catchment ranges from 600 m AOD (above ordnance datum) in the Pennine uplands, to below 20 m AOD in the Vale of York. Average annual rainfall varies with topography, reaching 2000 mm yr ÿ1 on the Pennine peaks and around 600 mm yr ÿ1 in low lying areas such as Selby to the south of York, at 5 m AOD. Daily mean and instantaneous maximum recorded discharges at Skelton (SE 425 734), 5 km upstream of York, are 48.6 m3 s ÿ1 and 622 m3 s ÿ1, respectively, with a mean annual ¯ood of 302 m3 s ÿ1. VARIATIONS IN FLOOD FREQUENCY AND MAGNITUDE IN THE YORKSHIRE OUSE 1878±1996 Previous ¯ood studies at York Floods at York have been documented as far back as the thirteenth century (e.g. Radley and Simms, 1971; Piers, 1977; Waterman, 1981), although much of this information is descriptive, non-systematic and biased towards recording of large, spatially extensive ¯oods (Archer, 1987; McEwen, 1987; Rumsby, 1991). The ®rst comprehensive investigation of ¯ood levels at York was commissioned by the Yorkshire Ouse River Board to examine an apparent increase in ¯ooding in several districts of York, and concluded, from analysis of ¯ood records between 1878 and 1951, that an increase in ¯ood frequency and magnitude had occurred during the 1940s and 1950s (Farrant, 1953). Little further research was conducted until catastrophic ¯ooding during the winter of 1981±1982 (cf. Lucas, 1982; Randon, 1982). This ¯ood, and an earlier event in December 1978, led to a feasibility study for the construction of a ¯ood barrier over the River Foss (see Miles, 1987) at its con¯uence with the Ouse at York (Mott, Hay & Anderson and Sir M. Macdonald & Partners, 1982). This report updated the work of Farrant (1953) and concluded that both ¯ood frequency and magnitude showed a statistically signi®cant upward trend. This recent increase in ¯ooding has been attributed to land use change upstream of York, including land drainage practices such as moorland gripping (Cau®eld, 1982; Old®eld, 1983) and overgrazing by sheep, resulting in a loss of vegetation cover and higher rates of runo€ (Sansom, 1996). Data sources for the study

Systematic ¯ood recording began in York in 1878 at Ouse Bridge (datum 4.97 m AOD) opposite the ¯oodprone King's Staithe area, and recorded all ¯oods above 8.03 m AOD (10 feet above normal summer level). Between 1893 and 1968 the Guildhall gaugeboard (datum 4.98 AOD) was used for recording, and from 1969 onwards a munro water level recorder and metric gaugeboard (datum 5.00 m AOD) has been in use, situated near the Viking Hotel, which has recently been replaced by a digital system. The fact that the sites are less than 200 m apart enables us to assume that the record is continuous and systematic since 1878. Stability of the channel cross-section at the stage recorders is dicult to assess, however, and between 1969 and 1982 the Yorkshire Water Authority (Mott, Hay & Anderson and Sir M. Macdonald & Partners, 1982) undertook hydrographic surveys of the River Ouse three times a year between Naburn and Poppleton. These surveys indicated that there have been no signi®cant changes in cross-sectional area over this period. Two ¯ood series are presented in this paper: annual maximum (AM) and peak-over-threshold (POT). The AM records represent the highest recorded ¯ood stage of each calendar year and provide a measure of ¯ood Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

1053

Figure 1. (a) Location of study area in the Yorkshire Ouse Basin. (b) Location of historic ¯ood recording sites in the city of York Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1054

S. A. LONGFIELD AND M. G. MACKLIN

magnitude. The POT series (all ¯oods above a speci®ed stage or discharge threshold) have been assessed for independence according to the rules de®ned by the NERC (1975) and a standard threshold adopted which gives an average of 4.5 ¯ood events per year over the 10-year period 1979±1988. Annual ¯ood frequency (AFF) series have been compiled from POT data by calculating the annual number of POT events over the standard threshold. Flood magnitude data have also been extracted from POT records through investigations of AFF at increasing stage thresholds based on return period estimates. The length of the ¯ood series at York is exceptional considering that the NERC (1995) calculated the average length of ¯ow record in the UK to be less than 23 years, and only 15 stations could be considered continuous for a period over 50 years. Only a small number of rivers in the UK have gauged ¯ood records that date to the nineteenth century, including the Lee at Fields Weir, the Thames at Teddington and the Severn Bewdley. Variability in annual maximum and peak-over-threshold records since 1878 Time-series plots of AM stage (1880±1996) and AFF (1878±1996) are presented in Figure 2, where longterm averages and ®ve-year running means are also plotted. The AM series at York show that ¯ood magnitudes in the period 1880±1903 were close to the long-term mean, with the exception of the large ¯ood of October 1892. There followed a period of extremely low ¯ood magnitudes in the early part of the twentieth century, until around 1915. From 1915 through to the present day there has been a series of

Figure 2. (a) Annual maximum ¯ood and (b) annual ¯ood frequency at York, 1878±1996. Dark line shows ®ve-year moving average, horizontal line shows long-term mean 1878±1996 Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

1055

15±20-year `saw-tooth' cycles with progressively increasing ¯ood stage, which peaked in the early 1980s. Within these cycles there are a series of peaks and troughs, with peaks in magnitude occurring in the late 1940s, mid to late 1960s, late 1970s to early 1980s and in the early 1990s. The highest ¯ood magnitudes over the whole period have occurred since 1977, with the two highest ®ve-year period averages being 1977±1981 (9.23 m AOD) and 1992±1996 (9.07 m AOD), compared with a long-term average of 8.53 m AOD. The ®veyear period with the lowest average ¯ood magnitude occurred between 1902 and 1906 (7.77 m AOD), with additional periods of low ¯ood magnitude recorded in the early 1930s, early 1950s and the early to mid1970s. Runs tests about the median con®rm that this as a non-random series with statistically signi®cant clustering of data (P value 5 005), indicating abrupt shifts between periods of high and low ¯ood magnitude. Kruskall±Wallis tests on ®ve-year period averages from 1880 also generate statistically signi®cant P values. The later part of the nineteenth century was characterized by generally low ¯ood frequencies, except between the late 1970s and early 1980s and the early 1940s ¯ood frequency remained extremely low, with three consecutive ®ve-year period averages between 1897 and 1911 having the lowest AFF in the entire record (average of 0.8 ¯ood events per year). Around 1944 there was a pronounced and sustained increase in ¯ood frequency. Large inter annual variations are clear in the 1950s ( ®ve-year running means staying close to the long-term average) until the mid-1960s, when there were several years with higher ¯ood frequencies. From around 1969 to 1977 there was a marked decline in ¯ood frequency, with the ®ve-year period between 1972 and 1976 giving an average of one ¯ood event per year, which is comparable with the lowest AFF recorded between 1897 and 1911. After 1977 the ¯ood frequency increased signi®cantly, with the periods 1977±1981 and 1982±1986 showing the highest average ®ve-year frequencies on record (4.20 events per year). Flood frequency since then has remained relatively high, apart from 1996. Runs and Kruskall±Wallis tests show that this is a non-random clustered series (P value 5 005). To evaluate the temporal distribution of larger ¯ood events, higher thresholds were calculated based on return period stage estimates derived using the POT exponential model (NERC, 1975). Figure 3 shows the annual frequency of ¯ood events at the standard threshold and the frequency of events exceeding the stage estimate of the two-year (Q2), 10-year (Q10) and 20-year ¯ood (Q20). All but one of the 13 events that have exceeded Q10 since 1878 have occurred after the 1920s, and four of the six ¯oods that have exceeded Q20 have occurred since 1978, indicating that there has been a marked increase in high magnitude ¯oods in recent years.

CONTROLS OF FLOODING AT YORK Climate variability The circulation of the atmosphere ultimately dictates the character of UK weather, and it is analysis of these circulation types that can most e€ectively explain and de®ne climatic variability (Kelly et al., 1997). Most previous studies that have investigated the relationships between atmospheric circulation and ¯ooding have inferred variations in circulation types from annual (Higgs, 1987; Rumsby and Macklin, 1994) or monthly (Knox et al., 1975) frequencies. Analysis of this type, however, gives only an indication of the general trends in climate that a€ect ¯ooding, and not the speci®c circulation types and weather conditions that generate individual ¯ood events. Grew (1996) has addressed this problem on Scottish rivers by classifying ¯ood events according to the circulation type on the day preceding the ¯oods and also on the ¯ood day. However, the number of days of precipitation that generate a ¯ood varies considerably depending upon catchment characteristics, particularly the size and gradient of the catchment (Grew, 1996). In this study a method has been developed using daily rainfall records to assess the key days responsible for ¯ood generation. The circulation type on these days can then be determined and a dominant circulation type established for each event. The chief advantage of this approach is that ¯oods are classi®ed according to the circulation type that provided the precipitation. Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1056

S. A. LONGFIELD AND M. G. MACKLIN

Figure 3. Floods on the River Ouse at York over increasing stage thresholds 1878±1996. Thresholds are calculated using the POT exponential model (NERC, 1975)

Method of analysis: ¯ood-generating circulation study. Daily rainfall records at Moorland Cottage (SD 807 923) and Blackmoorfoot (SE 096 130) were used to establish the speci®c days prior to each ¯ood at York that provided the majority of precipitation. These sites were chosen because of their proximity to the three main tributaries being considered and the length of record. Rainfall records four days before the ¯ood day were examined ( ®ve days in all including the day of ¯ood), and a minimum threshold of 10 mm rainfall per day was set, so as to consider only those days with a signi®cant volume of rain. Rainfall totals were then weighted using the following formula, which gives greater emphasis to those days with high rainfall totals Weighting ˆ …Daily rainfall ÿ Rainfall threshold† Copyright # 1999 John Wiley & Sons, Ltd.

2

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1057

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

Circulation types from the daily Lamb (1972) catalogue [see Kelly et al. (1997) for summary] on the relevant days prior to the ¯ood were then assigned the weighted value, and the type found to have the highest weighting was taken as the dominant synoptic type that generated that particular ¯ood event. Results of ¯ood-generating circulation study. Of the 252 POT ¯ood events that occurred at York between 1878 and 1995, 96% were classi®ed as being caused by a single circulation type. The percentage of ¯oods generated by each type is shown in rank order in Table I. A relatively small number of the 27 circulation types generated the majority of ¯oods at York: four circulation types, westerly, cyclonic, cyclonic-westerly and south-westerly generated 79.7% of all events. Westerly and cyclonic types are by far the most important ¯ood generators, accounting for 32.8% and 31.5% of all ¯oods, respectively (typical examples on 500 mb and surface weather charts are shown in Figure 4). Other circulation types that appear to be of signi®cant but secondary importance for ¯ood generation at York include southerly, cyclonic-south-westerly and cyclonicnortherly. To investigate whether certain circulation types generate ¯oods of a particular magnitude, the stage record at York was divided into magnitude ranges de®ned as follows Minor flood : Moderate flood : Major flood :

5 standard threshold 5 Q2 5 Q2 5 Q10 4 Q10

The results are displayed in Table I and highlight the dominance of westerly and cyclonic circulations at all magnitudes. Westerly circulations appear to be more dominant at moderate ¯ood magnitudes, generating 8.57% more ¯oods than cyclonics. However, the highest magnitude events (4Q10 ) are mostly generated by cyclonic circulations (41.7%). Table I. Rank order of the percentage of ¯oods generated by Lamb circulation types at York 1878±1996. Percentages are calculated for all POT ¯oods (4803 m AOD), and for minor, moderate and major magnitude ranges. Eight of the 27 Lamb types did not generate any ¯oods and have not been included Circulation type

% of total POT ¯oods generated

Westerly (W) Cyclonic (C) Cyclonic-westerly (CW) South-westerly (SW) Southerly (S) Cyclonic-south-westerly (CSW) Cyclonic-northerly (CN) Unclassi®ed (U) Cyclonic-easterly (CE) North-westerly (NW) Northerly (N) Anticyclonic (A) South-easterly (SE) North-easterly (NE) Easterly (E) Cyclonic-southerly (CS) Cyclonic-north-westerly (CNW) Anticyclonic-westerly (AW) Cyclonic-north-easterly (CNE) Copyright # 1999 John Wiley & Sons, Ltd.

32.78 31.54 9.54 5.81 3.73 2.9 2.49 2.49 1.66 1.24 1.24 0.83 0.83 0.83 0.41 0.41 0.41 0.41 0.41

% of ¯oods generated in magnitude ranges Minor (4standard threshold 5 Q2†

Moderate (4 ˆ Q2 5 Q10 )

Major (4Q10 )

31.45 31.45 9.43 5.66 3.77 2.52 3.14 2.52 1.89 1.26 1.26 1.26 1.26 Ð 0.63 0.63 0.63 0.63 0.63

38.57 30.00 11.43 5.71 2.86 2.86 1.43 1.43 Ð 1.43 1.43 Ð Ð 2.86 Ð Ð Ð Ð Ð

16.67 41.67 Ð 8.33 8.33 8.33 Ð 8.33 8.33 Ð Ð Ð Ð Ð Ð Ð Ð Ð Ð

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1058

S. A. LONGFIELD AND M. G. MACKLIN

Figure 4. Typical examples of cyclonic and westerly circulation types, shown on surface and 500 mb charts (after Lamb, 1972)

To determine whether changes in ¯ood frequency are related to circulation types the number of ¯ood events generated per year has been calculated over ®ve time periods, with a distinct ¯ood regime, for the four most common ¯ood-generating circulation types (Table II). In the period 1878±1943, westerly circulations generated more ¯oods per year than any other type. The trend continued throughout the period 1944±1968, which was characterized by a marked and sustained increase in ¯ood frequency. After this time the relative number of ¯oods generated under cyclonic conditions increased to surpass those generated under westerly conditions. This increase is particulary evident in the most recent period (1978±1996), which is characterized by high ¯ood frequency and increases in both south-westerly- and cyclonic-westerly-generated ¯oods. Weather associated with ¯ood-generating circulations. Owing to the considerable variability of depression tracks over long periods, cyclonic circulations tend to produce a fairly uniform distribution of precipitation over the British Isles (Sweeney and O'Hare, 1992). Westerly circulations are generally associated with unsettled or changeable weather, with the majority of rain falling in the northern and western parts of the British Isles (Lamb, 1972). Consequently, there is a marked west±east precipitation gradient associated with westerly type circulations, with a prominent rain shadow to the east of the Pennines (Sweeney and O'Hare, Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1059

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

Table II. Number of POT ¯ood events per year (4803 m AOD) generated by westerly, cyclonic, cyclonic-westerly and south-westerly Lamb circulation types over four time periods Period

1878±1903 1904±1943 1944±1968 1969±1977 1978±1996 Lois monitoring period 1993±1996

Westerly R Cyclonic R Cyclonic- R westerly

0.46 0.42 1.08 0.33 1.11 0.50

1 1 1 2 2 2

0.42 0.30 0.88 0.67 1.33 1.25

2 2 2 1 1 1

0.08 0.13 0.32 0.11 0.39 0.25

ˆ3 3 3 ˆ3 3 ˆ3

Southwesterly

R

Character of ¯ood frequency (FF) series

Character of ¯ood magnitude (FM) series

0.08 0.03 0.16 0.11 0.33 0.25

ˆ3 4 4 ˆ3 4

Low FF Low FF High FF Low FF High FF

Decreasing FM Increasing FM High and low FM Low FF High FM

ˆ3

High FF

High FM

R ˆ rank.

1992). Despite this it is clear that westerly circulation types play an important role in ¯ood generation in the Ouse Basin, which may re¯ect the fact that the majority of tributaries have headwaters in the Pennines, which receive high rainfall totals from westerly circulations. However, in the absence of a west±east precipitation gradient under cyclonic conditions, it is this weather type that delivers the highest amounts of precipitation over the whole Ouse Basin, and therefore generates some of the largest ¯oods. Cyclonic-westerly and southwesterly types have also been identi®ed as being important for ¯ood generation in the Ouse Basin. Cyclonicwesterlies tend to yield high rainfall totals and the west±east contrast is not as high as with pure westerly circulations, resulting in proportionately more rainfall falling in eastern parts of the country. The in¯uence of south-westerlies in ¯ood generation again relates to the fact that high precipitation totals are common with this type, although there is a marked west±east contrast in rainfall receipts. South-westerlies are warm, moist air masses and have been shown to trigger snowmelt ¯ood events at a number of locations (e.g. Jackson, 1978; Marsh and Monkhouse, 1991). Land use change Although climate is undoubtedly the primary `driver' of ¯ood events, changes in land use can cause signi®cant alterations in runo€ leading to variations in ¯ood frequency and magnitude. Upstream of York, major land use changes between 1878 and 1996 include land drainage, changes in agricultural practice, base metal mining and, to a lesser extent, construction of reservoirs and ¯ood defence schemes. Many of these can also in¯uence sediment availability, though historic metal mining in the Pennine ore®elds, as elsewhere in Britain, has been shown to be of primary importance with respect to sediment-associated contaminant metals (Macklin et al., 1994). Lead mining in the Ouse Basin was concentrated in the Upper Swale, Ure, Nidd and Wharfe areas. Large-scale production increased markedly towards the end of the eighteenth century and peaked in the mid-nineteenth century (Fieldhouse and Jennings, 1978; Macklin et al., 1997). The ¯ood record at York dates from 1878, when ore production was in decline, though even after this period historic metal mining is one of the principal controls of sediment and contaminant supply because of remobilization of metal-polluted alluvium from ¯oodplains during large ¯ood events (Macklin and Dowsett, 1989). Land drainage schemes have also had an e€ect on the ¯ood regime. In lowland areas, land drainage consists of a system of subsurface pipes (underdrainage) which can be combined with `secondary' treatments such as moling and subsoiling. In upland areas, pipe systems are not economical and drainage usually consists of a system of open ditches (Robinson, 1990). Much interest has focused on the e€ects of hill drainage or `moorland gripping' on ¯ood peaks and large-scale upland erosion in the Ouse Basin, particularly following serious ¯ooding in the winter of 1981±1982, when it was suggested that drainage schemes may have exacerbated downstream ¯ooding (Cau®eld, 1982; Old®eld, 1983). In the Yorkshire Dales the most Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1060

S. A. LONGFIELD AND M. G. MACKLIN

Figure 5. Percentage of land upstream of York that has had moorland gripping (after Robinson, 1990)

heavily gripped areas (450% of the land) are in the upper reaches of the River Nidd and in Arkengarthdale (see Figure 5) a tributary of the River Swale, though the headwaters of all of the major tributaries above York have been subject to gripping to some extent (Robinson, 1990). Large-scale gripping was initiated in the Yorkshire Dales in the 1940s, prompted by the introduction of grant-aid and a greater need for livestock production during the Second World War, which continued until the 1960s. Rates of upland drainage increased markedly in the 1970s and declined in the 1980s to very low levels at present (Robinson, 1990). Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

1061

The most intensive period of lowland underdrainage occurred over the 30-year period between 1840 and 1869 (Phillips, 1989), pre-dating the York gauged ¯ood record. There followed a prolonged period of agricultural depression between the 1880s and 1930s, which resulted in a marked reduction in the number of drainage schemes (Nicholson, 1943; Tra€ord, 1970). At the onset of the Second World War, in an e€ort to increase agricultural production, government grant-aid was introduced in England and Wales towards the cost of underdrainage, hill drainage and arterial channel drainage (Robinson, 1990). The area drained annually in England and Wales increased from 1939 and reached a peak of around 100 000 hectares per year in the 1970s (Stans®eld, 1987; Robinson, 1990). In the Yorkshire Ouse Basin, the majority of underdrainage has been concentrated in the southern part of the Vale of York. No statistics are available after 1980, although it is generally accepted that the level of drainage activity has declined since the peak of the 1970s (Stans®eld, 1987). The third major land use change to consider is variation in agricultural practices, which can alter rates of runo€ and sediment production. Heathwaite et al. (1990) used experimental plot studies to examine the e€ects of di€erent land use types on soil in®ltration capacity, runo€, bulk density, sediment yield and sediment delivery rates. Highest runo€ volumes and sediment yields were recorded for heavily grazed permanent pasture, owing to soil compaction and stripping of vegetation cover through overgrazing. Runo€ and sediment yields were also high from bare ground (i.e. with no crops), even though land had recently been ploughed and rolled. Lowest runo€ volumes and sediment yields were recorded on ungrazed land (i.e. temporary grassland and cereals), where higher in®ltration capacity and lower soil bulk density resulted in markedly less rainfall being converted to runo€. Annual agricultural returns from 1866 to 1939 have previously been compiled for the North Riding of Yorkshire by Wooldridge (1945). this series was updated from 1939 to 1996 using the annual publication of the Agricultural Statistics, UK. Two main land classi®cations were used between 1866 and 1980, total area of arable crops and total area of permanent grassland, which together represent the principal land uses in the Yorkshire Ouse Basis upstream of York. Areas of arable crops and permanent grassland are shown in Figure 6 on a yearly basis for the North Riding of Yorkshire (the majority of land upstream of York was in the North Riding prior to county boundary reorganization in 1974). Between 1866 and 1914 the total arable area declined by 6.6%, whilst over the same period permanent grassland increased by 13.8%. The increased need for home food production in the First World War temporarily reversed this trend (Sheail, 1973) and resulted in a 3% increase in arable area between 1914 and 1918. Following the end of the First World War a period of economic depression caused large areas of arable land to revert back to grassland (Bowler, 1991) with arable area decreasing by 6.1% between 1918 and 1939. The most dramatic shift in agricultural land use

Figure 6. Annual area under arable crops and permanent grassland in the North Riding of Yorkshire 1866±1973, and for North Yorkshire 1974±1996 in arbitrary units to take account of county boundary reorganization Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1062

S. A. LONGFIELD AND M. G. MACKLIN

occurred at the onset of the Second World War. Increasing government grants and a sustained `plough-up' campaign resulted in a 12.3% increase in arable area in the North Riding, associated with a 14.7% decline in the area of permanent grassland. Immediately after the Second World War arable cultivation slightly decreased, but since the mid-1950s has increased, while the area of permanent grassland has declined. Furthermore, sheep numbers have increased by nearly 40% since 1982 (Sansom, 1996), and it has been suggested that this may have contributed to recent severe ¯ooding, particularly in Wensleydale, by promoting more rapid runo€ through loss of vegetation cover and poaching of the soil. Both reservoirs and ¯ood defence schemes have had little in¯uence on variations in ¯ood regime or sediment supply. Only 4.71% of the catchment upstream of York has been impounded and there are few major ¯ood defence schemes upstream of York, with the exception of the large washland at Clifton Ings and defence schemes around Boroughbridge. CHANGES IN FLOOD FREQUENCY AND MAGNITUDE AND ITS LIKELY EFFECT ON SEDIMENT FLUXES One of the research questions we sought to address was how changes in ¯ood frequency and magnitude have a€ected long-term ¯uxes of ®ne sediment and associated contaminants. To evaluate this, and to assess how representative the LOIS monitoring period has been in terms of ¯ood regime and sediment ¯uxes over the last 100 years or so, three time periods are considered: (1) 1878±1996, (2) the LOIS monitoring period, 1993± 1996, and (3) forecasts for the next 50 years or so. For this qualitative reconstruction of sediment and contaminant ¯uxes we assume, on the basis of ®eld observations made in the regions rivers during (Wass et al., 1997) and before (Passmore et al., 1993; Rumsby and Macklin, 1994) the LOIS programme, that major ¯oods result in large-scale bank erosion and the remobilization of contaminated ¯oodplain material (especially in upland and piedmont reaches). While the e€ects of moderate ¯oods are primarily restricted to ¯ushing of sediments stored within the channel, this does not, however, assume a simple linear relationship between sediment ¯ux and ¯ood frequency and magnitude, since sediment availability is strongly in¯uenced by land use changes. This section simply attempts to place the LOIS monitoring period in a longer term context based on hydrology and land use changes. 1878±1996 The ¯ood record between 1878 and 1996 has been sub-divided into ®ve periods, each having a distinct ¯ood regime (Table II). 1878±1903 The end of the nineteenth century and the beginning of the present century were characterized by relatively low ¯ood frequencies and magnitudes close to the long-term means, with only one event exceeding Q20 . Westerly circulations were the dominant ¯ood generator, closely followed by cyclonics. Fine sediment supply was high in the Yorkshire Dales uplands owing to base metal mining (Macklin et al., 1997), though in the lowlands reversion of arable to grassland is likely to have reduced sediment availability. This, combined with low ¯ood frequency and magnitude, would have resulted in relatively low sediment ¯uxes, but as a consequence of mining activity, metal contaminant concentrations and downstream delivery would have been high. 1904±1943 Flood frequency further declined and magnitude remained very low, with the majority of ¯oods generated under westerly conditions. The cessation of base metal mining towards the end of the nineteenth century would have resulted in a marked reduction in contaminant metal and ®ne sediment supply. There were no other major land use changes during this period (although a number of major changes were initiated at the beginning of the Second World War) and it is likely that both sediment and contaminant ¯uxes were low, probably the lowest in the 119-year record. Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1063

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

1944±1968 The 1940s marks the onset of very signi®cant changes in land use, climate and ¯ood regime. There was a marked and sustained increase in ¯ood frequency, with high ¯ood magnitudes in both the late-1940s and the mid±late 1960s. Floods were generated predominately by westerly circulations, despite a decline in annual frequencies of westerlies since the 1950s (Bri€a et al., 1990; Jones and Kelly, 1982). Increasing government grant-aid resulted in a major expansion of moorland gripping in the Yorkshire Dales. There was also an increase in lowland underdrainage after 1939, associated with a massive plough-up campaign to meet the demand for increased home food production during the Second World War. Increased sediment supply from both upland gripping and lowland cereal cultivation, together with higher ¯ood frequencies, would have signi®cantly increased sediment ¯uxes in comparison with the early decades of the twentieth century. 1969±1977 For the ®rst time cyclonic circulations became the dominant ¯ood generator, with ¯ood frequency and magnitude falling sharply over this 10-year period. Large-scale drainage in both upland and lowland areas continued, although extremely low ¯ood frequencies and ¯ood magnitudes would have resulted in considerably reduced sediment ¯uxes. 1978±1996 The most recent period has experienced the highest ¯ood frequencies and magnitudes in the last 119-years, with four events exceeding Q20 . Cyclonic weather types generated the majority of ¯oods, re¯ected in a dramatic increase in the annual frequency of cyclonic circulations over the past two decades (Jones and Kelly, 1982). The number of south-westerly-generated ¯oods has also increased, associated with a rise in the annual frequency of southerly air¯ows (Murray, 1993) and south-westerly circulations (Sweeney and O'Hare, 1992) since the 1960s. In terms of ®ne sediment transport, four very large ¯ood events (1978, 1982, 1991 and 1995) are likely to have remobilized considerable volumes of ¯oodplain sediment through bank erosion. Sediment supply is also likely to have increased markedly in this period, associated with rapidly increasing numbers of grazing animals and a switch to winter cereals (cf. Evans and Cook, 1986; Boardman, 1990, 1995; Boardman et al., 1994), both of which probably enhanced sediment production. The last two decades has probably experienced the highest rates of sediment ¯ux over the past 120 years and with largescale remobilization of mining-contaminated alluvium, notably during the 1986 (Newson and Macklin, 1990) and 1995 ¯ood events. Indeed, contaminant metal mineral loads may be approaching those of the late nineteenth century. The LOIS monitoring period, 1993±1996 Table III shows that the LOIS monitoring period has been characterized by higher than average AM ¯oods and POT events per year. The majority of ¯oods were generated by cyclonic circulations, which have been shown in earlier parts of the record to be associated with high magnitude events and high rates of Table III. Summary of ¯ood characteristics in the historical period and during the LOIS monitoring programme Flood characteristics Average AM ¯ood Max. AM Flood Min. AM ¯ood Average no. of POT ¯ood events per year Max. no. of POT ¯ood events per year Min. no of POT ¯ood events per year Copyright # 1999 John Wiley & Sons, Ltd.

1878±1992

1878±1996

1993±1996 (LOIS monitoring period)

8.52 10.12 (Jan 1982) 6.96 (Nov 1905) 2.09 9 0

8.53 10.12 (Jan 1982) 6.96 (Nov 1905) 2.14 9 0

8.85 9.67 (Feb 1995) 8.32 (Feb 1996) 3.60 5 2

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1064

S. A. LONGFIELD AND M. G. MACKLIN

sediment ¯ux. The largest ¯ood occurred in February 1995, which was ranked ®fth in the 119-year record. Field observations during and following the event highlighted signi®cant bank erosion and remobilization of contaminated ¯oodplain sediment. The LOIS monitoring period has therefore been characterized by very high ¯ood frequency and magnitude and, in an historical context, signi®cantly enhanced sediment and contaminant metal ¯ux. Future trends in ¯ood frequency, magnitude and sediment ¯uxes Recent research in the USA (Knox, 1993), Australia (Smith, 1993) and Britain (Beven, 1993) suggests that future climate change associated with global warming will cause marked changes in ¯ood frequency and magnitude. Climate change scenarios for the British Isles give a general consensus that with rising temperatures precipitation will increase in winter (Rowntree, 1990; UKCCIPG, 1991; Arnell, 1992; Rowntree et al., 1993; Raper et al., 1997) and that this will be greatest in the north and west (Santer et al., 1990; Arnell, 1992). The most probable e€ect of postulated increases in winter rainfall, storminess and ¯ood-producing rainfalls in the Ouse Basin would be a continued increase in ¯ood frequency and magnitude. However, rising temperatures may reduce snow cover, which could have a signi®cant e€ect on ¯ood magnitude since most of the largest historical ¯ood events in the Yorkshire Ouse Basin have been in¯uenced by snowmelt. Assessing the future e€ects of changes in the atmospheric circulation on ¯ooding is, however, more problematic. Rumsby and Macklin (1994) suggest that an increase in global temperature would reduce the equator±pole temperature gradient and lead to an enhanced zonal circulation regime, which would in turn increase the frequency of westerly circulation types. Alternatively, Sweeney and O'Hare (1992) and Jones (1992) suggest that a reduced thermal gradient would result in a continued decline in westerly circulations, which could result in a higher frequency of large ¯ood events because of increased convective activity and more intense depressions. In terms of sediment dynamics, any increase in cyclonic activity would result in a higher frequency of extreme ¯ood events and increased sediment ¯uxes. CONCLUSIONS Analysis of River Ouse ¯ood records at York has demonstrated that there have been signi®cant variations in ¯ood frequency and magnitude over the past 120 years. Flood events have been found to be generated by a relatively small number of atmospheric circulation conditions, with westerly and cyclonic types being the most important. Generally, westerly air¯ows produce moderate ranking ¯oods, whereas extreme ¯oods are more commonly associated with cyclonic circulations. Periods characterized by low ¯ood magnitude have a higher frequency of westerly air¯ows, whilst periods with the highest ¯ood magnitudes are dominated by cyclonic-generated ¯oods. From ¯ood and land use records it appears that sediment ¯uxes were probably relatively low from the end of the late nineteenth century to the mid-1940s. This resulted from low ¯ood frequency and magnitude and reduced sediment supply following the cessation of base metal mining. Largescale land use changes beginning in the Second World War signi®cantly increased sediment supply and, together with an increase in ¯ood frequency, is likely to have resulted in higher sediment ¯uxes. The highest ¯ood magnitudes and frequencies in the 119-year stage record, however, have occurred since 1978. This period encompasses the LOIS monitoring programme and has been associated with an increase in the frequency of cyclonic circulations and cyclonic-generated ¯oods. Four of the six largest ¯ood events on record have occurred over this period and sediment ¯uxes are likely to have been highest since the end of the nineteenth century. ACKNOWLEDGEMENTS

The authors would like to thank NERC for funding LOEPS Special Topic 12 within the framework of the Land±Ocean Interaction Study (LOIS). Thanks also to Ian Ward and Mark Fuller of the Environment Agency at York for providing access to the historical ¯ood record at York, to Alison Manson for producing Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

THE INFLUENCE OF RECENT ENVIRONMENTAL CHANGE

1065

some of the ®gures and to the anonymous referee for constructive comments. This paper represents publication number 352 of the LOIS project.

REFERENCES

Archer, D. 1987. `Improvement in ¯ood estimates using historical ¯ood information on the River Wear at Durham'. in Proceedings of the National Hydrology Symposium. British Hydrological Society, pp. 5.1±5.9. Arnell, N. W. 1992. `Impacts of climatic change on river ¯ow regimes in the UK', J. Inst. Wat. Environ. Mgmt., 6, 432±442. Beven, K. 1993. `Riverine ¯ooding in a warmer Britain', Geogr. J., 159, 157±161. Boardman, J. 1990. `Soil erosion of the South Downs: a review', in Boardman, J., Foster, I. D. L. and Dearing, J. A. (Eds), Soil Erosion of Agricultural Land. John Wiley and Sons Ltd, Chichester, pp. 87±105. Boardman, J. 1995. `Damage to property by runo€ from agricultural land, South Downs, southern England', Geogr. J., 161, 177±191. Boardman, J., Ligneau, L., de Roo, A., and Vandaele, K. 1994. `Flooding of property by runo€ from agricultural land in northwestern Europe', Geomorphology, 10, 183±196. Bowler, I. R. 1991. `The agricultural pattern', in Johnson, R. J. and Gardiner, V. (Eds), The Changing Geography of the United Kingdom. Routledge, London, pp. 83±114. Bri€a, K. R., Jones, P. D., and Kelly, P. M. 1990. `Principal component analysis of the Lamb catalogue of daily weather types: part 2, seasonal frequencies and update to 1987', Int. J. Climatol., 10, 549±563. Cau®eld, C. 1982. `Too much damage spells trouble on t'farm', New Scientist, 159, 125. Evans, R. and Cook, S. 1986. `Soil erosion in Britain', SEESOIL, 3, 28±58. Farrant, A. E. 1953. `A Memorandum on Flooding at York'. Unpublished report to the Yorkshire Ouse River Board, 64 pp. Fieldhouse, R. and Jennings, B. 1978. A History of Richmond and Swaledale. Phillmore & Co, London. Fox, I. A. and Johnson, R. C. 1997. `The hydrology of the River Tweed', Sci. Total Environ., 194/195, 163±172. Grew, H. L. 1996. `Temporal variability in the ¯ooding of Scottish rivers'. PhD Thesis, University of St Andrews. 2 volumes. Heathwaite, A. L., Burt, T. P., and Trudgill, S. T. 1990. `Land-use controls on sediment production in a lowland catchment, south-west England', in Boardman, J., Foster, I. D. L. and Dearing, J. A. (Eds), Soil Erosion on Agricultural Land. John Wiley and Sons Ltd, Chichester, pp. 69±86. Higgs, G. 1987. `Environmental change and ¯ood hazard in the Upper River Severn'. PhD Thesis, University of Wales, Aberystwyth, 574 pp. Jackson, M. C. 1978. `The in¯uence of snowmelt on ¯ood ¯ows in rivers', J. Inst. Wat. Eng., 32, 495±508. Jarvie, H. P., Neal, C., and Robson, A. J. 1997. `The geography of the Humber catchment', Sci. Total Environ., 194/195, 87±99. Jones, P. D. 1992. `The projected impact of global warming on hydrological regimes in western Europe', in Changming, L and Guobin, F. (Eds), Regional Hydrological Responses to Climatic Change and Global Warming. Internat. Geogr. Union, Washington D.C., pp. 25±33. Jones, P. D. and Kelly, P. M. 1982. `Principal component analysis of the Lamb catalogue of daily weather types: part 1, annual frequencies', J. Climatol., 2, 147±157. Kelly, P. M., Jones, P., and Bri€a, K. 1997. `Classifying the winds and weather', in Hulme, M. and Barrow, E. (Eds), Climates of the British Isles Ð Present, Past and Future. Routledge, London, pp. 153±172. Knox, J. C. 1993. `Large increases in ¯ood magnitude in response to modest changes in climate', Nature, 361, 430±432. Knox, J. C., Bartlein, P. J., Hirschboeck, K. K., and Muckenhirn, R. J. 1975. `The response of ¯oods and sediment yields to climate and land-use in the Upper Mississippi Valley', Univ. Wis. Inst. Environ. Stud. Rep., 52, 1±76. Lamb, H. H. 1972. `British Isles weather types and a register of the daily sequence of circulation patterns 1861±1971', Meteorological Oce Geophysical Memoir. 116, HMSO, London. Law, M., Wass, P., and Grimshaw, D. 1997. `The hydrology of the Humber Catchment', Sci. Total Environ., 194/195, 119±128. Lucas, S. 1982. `Winter of frosts, ¯oods and bursts', Water, 43, 2±5. Macklin, M. G. and Dowsett, R. B. 1989. `The chemical and physical speciation of trace metals in ®ne grained overbank ¯ood sediments in the Tyne basin, north-east England', Catena, 16, 135±151. Macklin, M. G., Ridgway, J., Passmore, D. G., and Rumsby, B. T. 1994. `The use of overbank sediment for geochemical mapping and contaminant assessment: results from selected English and Welsh ¯oodplains', Appl. Geochem., 9, 689±700. Macklin, M. G., Hudson-Edwards, K. A., and Dawson, E. J. 1997. `The signi®cance of pollution from historic metal mining in the Pennine ore®elds on river sediment contaminant ¯uxes to the North Sea', Sci. Total Environ., 194/195, 391±197. Marsh, T. J. and Monkhouse, R. A. 1991. `A year of hydrological extremes Ð 1990', Weather, 46, 366±376. Marsh, T. J. and Sanderson, F. J. 1997. `A review of hydrological conditions throughout the period of the LOIS monitoring programme Ð considered within the context of the recent UK climatic variability', Sci. Total Environ., 194/195, 59±69. McEwen, L. J. 1987. `Sources for establishing a historic ¯ood chronology ( pre-1940) within Scottish river catchments', Scott. Geogr. Mag., 103, 132±140. Miles, L. 1987. `Fighting the ¯oods: York barrier falls into place', Surveyor. 1 October. Mott, Hay & Anderson in association with Sir M. Macdonald & Partners, 1982. River Foss Flood Alleviation Ð Draft Report. Yorkshire Water Authority, Leeds. Murray, R. 1993. `Bias in Southerly types in decade 1981±90 over the British Isles', Weather, 48, 152±154. NERC 1975. The Flood Studies Report. 5 volumes, Natural Environmental Research Council, London. NERC 1995. Hydrological Data United Kingdom 1994 Yearbook. Institute of Hydrology, Wallingford, 176 pp. NERC 1996. Hydrological Data United Kingdom 1995 Yearbook. Institute of Hydrology, Wallingford, 176 pp. Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)

1066

S. A. LONGFIELD AND M. G. MACKLIN

Newson, M. D. and Macklin, M. G. 1990. `The geomorphologically-e€ective ¯ood and vertical instability in river channels Ð a feedback mechanism in the ¯ood series for gravel-bed rivers' in White W. R. (Ed.), International Conference on River Flood Hydraulics. Wiley and Sons, Chichester, pp. 123±140. Nicholson, H. H. 1943. `Modern ®eld drainage', J. R. Agric. Soc. Engl., 104, 118±135. Old®eld, P. 1983. `Drain and be damned', The Guardian. 9th April. Passmore, D. G., Macklin, M. G., Brewer, P. A., Lewin, J., Rumsby, B. T., and Newson, M. D. 1993. `Variability of late Holocene braiding in Britain', in Best, J. L. and Bristow, C. S. (Eds), Braided Rivers. Geological Society Special Publication No. 75, pp. 205±229. Phillips, A. D. M. 1989. The Underdraining of Farmland in England during the Nineteenth Century. Cambridge University Press, Cambridge, 315 pp. Piers, T. 1977. `And the waters increased', Yorkshire Life, 1977, 30±33. Radley, J. and Simms, C. 1971. Yorkshire Flooding Ð Some E€ects on Man and Nature. Ebor Press, York, 28 pp. Randon, D. V. 1982. `Floods in the York and Selby area, January 1982', J. Meteorol., 7, 41±43. Raper, S., Viner, D., Hulme, M., and Barrow, E. 1997. `Global warming and the British Isles', in Hulme, M. and Barrow, E. (Eds), Climates of the British Isles Ð Past and Future. Routledge, London, pp. 326±339. Robinson, M. 1990. Impact of Improved Land Drainage on River Flows. Institute of Hydrology Report No. 113, Institute of Hydrology, Wallingford, 226 pp. Rowntree, P. R. 1990. `Estimates of future climatic change over Britain. part 2: results', Weather, 45, 79±89. Rowntree, P. R., Murphy, J. M., and Mitchell, J. F. B. 1993. `Climatic change and future rainfall predictions', J. Inst. Wat. Environ. Mgmt, 7, 464±470. Rumsby, B. T. 1991. `Flood frequency and magnitude estimates based on valley ¯oor morphology and ¯oodplain sedimentary sequences: The Tyne Basin, N.E. England'. PhD Thesis, University of Newcastle Upon Tyne, 205 pp. Rumsby, B. T. and Macklin, M. G. 1994. `Channel and ¯oodplain response to recent abrupt climate change: The Tyne basin, Northern England', Earth Surf. Process. Landf., 19, 499±515. Sansom, A. 1996. `Floods and sheep Ð is there a link?', Circulation, 49, 1±4. Santer, B. D., Wigley, T. M. L., Schlesinger, M. E., and Mitchell, J. F. B. 1990. Developing Climate Scenarios from Equilibrium GCM Results. Max Planck Institute fuÈr Meteorologie, Report No. 47, Max Planck Institute, Hamburg. Sheail, J. 1973. `Changes in the use and management of farmland in England and Wales 1915±19', Trans. Inst. Br. Geog., 60, 17±32. Smith, D. J. 1993. `Greenhouse climatic change and ¯ood damages, the implications', Climatic Change, 25, 319±333. Stans®eld, C. 1987. `Field drainage 1940±1984 Ð a success! But where to now?', Soil Water, 15, 12±13. Sweeney, J. C. and O'Hare, G. P. 1992. `Geographical variations in precipitation yields and circulation types in Britain and Ireland', Trans. Inst. Br. Geog., 17, 448±463. Tra€ord, B. D. 1970. `Field drainage', J. R. Agric. Soc. Engl., 131, 129±152. UKCCIRG, 1991. The Potential E€ects of Climate Change in the United Kingdom, UK Climate Change Impacts Review Group, First Report, HMSO, London. Wass, P. D., Marks, S. D., Finch, J. W., Leeks, G. J. L., and Ingram, J. K. 1997. `Monitoring and preliminary interpretation of in-river turbidity and remote sensed imagery for suspended sediment transport studies in the Humber catchment', Sci. Total Environ., 194/195, 263±283. Waterman, R. D. 1981. `Flooding in the city of York', Chart. Mun. Eng., 108, 257±261. Webb, B. W., Phillips, J. M., Walling, D. E., Littlewood, I. G., Watts, C. D., and Leeks, G. J. L. 1997. `Load estimation methodologies for British rivers and their relevance to the LOIS RACS(R) programme', Sci. Total Environ., 194/195, 379±389. Wooldridge, S. W. 1945. in Stamp L. D. (Ed.), The Land of Britain, The Report of the Land Utilisation Survey of Britain, Part 51, Yorkshire (North Riding), Geographical Publications Limited, London.

Copyright # 1999 John Wiley & Sons, Ltd.

HYDROLOGICAL PROCESSES, VOL. 13, 1051±1066 (1999)