THE IMPACT OF LATE HOLOCENE CLIMATIC VARIABILITY AND LAND-USE CHANGE ON FLOOD HYDROLOGY (GUADALENTIN RIVER,

SE SPAIN)

G. Benito1, M. Rico2, Y. Sánchez-Moya3, A. Sopeña3, V. R. Thorndycraft4 and Barriendos, M.5

1Institute of Natural Resources, Consejo Superior de Investigaciones Científicas,

Serrano 115 bis, 28006 Madrid, Spain. 2Instituto Pirenaico de Ecología, CSIC, Zaragoza, Spain

3 Instituto de Geología Económica, CSIC-Universidad Complutense, 28040 Madrid,

Spain 4Department of Geography, Royal Holloway, University of London, Egham, Surrey,

TW20 0EX, U.K. 5Department of Modern History, University of Barcelona, B. Reixac s/n, 08028

Barcelona, Spain.

Abstract

The Guadalentín River, located in SE Spain, is considered one of the most torrential

rivers in Spain, as indicated by catastrophic events such as the 1879 flood that caused

777 fatalities in the Murcia region. In this paper, flood frequency and magnitude of

the upper Guadalentín River was reconstructed using geomorphological evidence,

combined with one dimensional hydraulic modelling and supported by records from

documentary sources at Lorca in the lower Guadalentin catchment. Palaeoflood

studies were conducted at a 2.5 km reach located at the confluence of the Rambla

Mayor (162 km2) and Caramel River (210 km2). These tributaries join at the entrance

of a narrow bedrock canyon, carved in Cretaceous limestone, which is 15-30 m wide

and 40 m deep. Six stratigraphic profiles were described, the thickest and most

complete corresponding to flood benches deposited upstream of the canyon

constriction. The stratigraphic and documentary records identify five main phases of

increased flood frequency. Phase 1, based on sedimentary palaeoflood evidence

alone, occurred at ca. 950-1200 cal AD with at least ten floods with minimum

discharge estimates of 15-580 m3s-1. Phases 2-5, identified through combined

sedimentary and documentary evidence occurred at: (a) AD 1648-1672, with eight

documentary floods and two palaeofloods exceeding 580-680 m3s-1; (b) AD 1769-

1802, comprising seven documentary floods, of which at least two events (>250 m3s-

1) are preserved in the sedimentary record; (c) AD 1830-1840, with four documentary

floods, and at least two events recorded in the stratigraphy (760-1035 m3s-1); and

finally (d) the AD 1877-1900 period that witnessed seven documentary floods, with

three palaeofloods (>880 m3s-1). The palaeoflood and historical flood information

indicate an anomalous increase in the frequency of large magnitude floods between

AD1830-1900, which can be attributed to climatic variability accentuated by

intensive deforestation and land-use practices during the first decades of the 19th

century.

Key words: Palaeoflood hydrology, documentary records, floods, paleodischarges,

Little Ice Age, environmental change, SE Spain,

1. Introduction The assessment of the response of hydroclimatic extremes to anthropogenic global

change is one of the main future uncertainties according to the latest IPCC report (in

Trenberth et al., 2007). The understanding of extreme event patterns in the context of

global change is crucial due to its social (e.g. vulnerability) and political (resilience

and adaptation) implications, in particular for those regions where extremes are an

intrinsic characteristic of the hydrological regime. This especially applies to the

Mediterranean as it is characterised by a fragile hydrologic and environmental

equilibrium linked to frequent droughts and/or severe flooding (Vita-Finzi, 1969;

Butzer, 1974, 1975; Lewin et al., 1995). Long-term changes of rare events are,

however, difficult to identify due to short gauge records and their limited spatial

distribution (Frei and Schär, 2001; Klein Tank and Können, 2003). These

instrumental records can be lengthened by hundreds to thousands of years by

reconstructing discharges of past floods using palaeoflood evidence (e.g. Baker et al.,

1983) and/or written descriptions of floods using historical, archival documents (see

Brazdil et al., 2006). Slackwater palaeoflood sediments (Baker et al., 2002), deposited

in flow marginal areas of bedrock canyons during high flood stages, have been

employed in many regions of the world for compiling long term flood magnitude and

frequencies of large floods over decades to millennia, with an emphasis on

understanding extreme event response to Holocene climatic variability (e.g. Ely et al.,

1993; Benito et al., 2003a; Sheffer et al., 2007; Thorndycraft et al., 2006).

During the Holocene, however, hydrological sensitivity was driven not only by

climatic variability (Ely, 1997; Gregory et al., 2006) but also by human activities

(Foulds and Macklin, 2006). The geographic location of south-east Spain at the

southern limit of influence of the polar front in the tropically sensitive Mediterranean

zone provides great potential for river systems to reflect short-term regional climatic

variability e.g. through changes in flood frequency and magnitude. Furthermore, the

region has a long history of anthropogenic land-use change, which may lead to

increasing rates of sediment yield and changes in flood magnitude and frequency (see

Knighton, 1998, pp.316-322). Dealing with Holocene fluvial sedimentary records it

is often difficult to disentangle the role played by climate and humans in driving the

changes observed in river channels and floodplain sediments (Foulds and Macklin,

2006). In contrast to alluvial floodplain depositional environments where

sedimentation may result from a range of overbank flood magnitudes, slackwater

flood deposits are usually emplaced by high stage floodwaters caused by extreme

rainfall events whose frequency can be related to changing climate conditions

(Hirsboeck, 1991; Benito et al., 2003a), whilst changes in sedimentological

characteristics may be linked to environmental changes at the basin scale

(Thorndycraft et al., 2004). A major limitation, however, in establishing the spatial

and temporal occurrence of these palaeoflood records are the uncertainties inherent in

geochronological techniques, such as measurement error and calibration in the case of

radiocarbon dating (Trumbore, 2000). The combination of palaeoflood and

documentary data sources has the potential to reduce these uncertainties and provide a

more holistic approach that can strengthen our understanding of past floods and help

elucidate the effect of climate and land-use changes on flood hydrology (Benito et al.,

2004, Thorndycraft et al., 2006). This study contributes to this debate through the

reconstruction of a 1000 year flood record from the Guadalentín River basin (SE

Spain), In this paper we highlight the potential of combined sedimentological and

archival research into past floods of European river basins. The main objectives are

to: (1) reconstruct centennial scale flood frequency using the stratigraphic record of

slackwater flood deposits; (2) analyse the palaeoflow hydraulics associated with these

floods and calculate estimates of peak flood discharges; (3) apply the historical flood

information compiled from archival sources to refine the palaeoflood chronology; and

(4) synthesise the collected date to gain a detailed understanding of changing flood

magnitude and frequency during the last millennium, and in particular to document

flood response to regional climatic and local environmental changes.

2. Study area The Guadalentín River (the name deriving from the Arabic Oued al Lentin meaning

river of mud) is located in SE Spain and is a major tributary of the Segura River, one

of the largest Mediterranean basins in Spain (Fig.1). The Guadalentín River is one of

the most torrential rivers in the Mediterranean region (Pardé, 1961) with over 2000

fatalities documented during the last 250 years (Muñoz, 1989; Bautista and Muñoz,

1986), including a dam break event that killed over 600 people in AD1802. The mean

annual precipitation ranges from over 1000 mm in the high mountains of the Sierra de

Maria to less than 300 mm in the lower catchment. The region is affected by: (a)

Atlantic frontal systems (zonal flow) during late winter (February-March) and spring

(May-June), when the highest annual precipitation values are registered; and (b)

autumnal (September to November) mesoscale convective systems born from

southwest flows at altitude. The latter generate the majority of floods in the region

(Capel, 1981; Benito et al., 1996). For example, in October 1973 convective storm

cells produced a daily rainfall in excess of 100 mm in Lorca and of 180 mm in the

Guadalentín headwaters (Capel, 1974). This rainfall, together with the steep gradients

in the catchment, resulted in severe flash flooding registering a peak discharge of

3544 m3s-1 at the Puentes reservoir. The event caused 83 fatalities in Puerto

Lumbreras and 13 in Lorca (Fig. 1). The nearest gauge station to the study area is

located at the Valdeinfierno dam and drains an area of 456 km2. The gauge record

covers the periods 1933-1949 and 1968 onwards, with the largest recorded daily flood

discharge being 178 m3s-1 in 1946, though it is important to note that most of this

gauge record corresponds to mean daily discharges. The most reliable gauge station is

located at the Puentes reservoir (15 km downstream of the study site) where the

catchment size is substantially increased to 1428 km2 by the inflow of the Corneros

(552 km2) and Turrillas (340 km2) rivers.

The palaeoflood study site (Fig. 1) was centred on a 2.5 km reach in the upper

catchment at the confluence of the Rambla Mayor (162 km2) and Caramel River (210

km2), both ephemeral streams. The geology is dominated by Mesozoic limestone and

marls of the Sub-Betic domain and the confluence of the two tributaries is located at

the entrance of a narrow bedrock canyon carved in Cretaceous limestones. The gorge

is 15-30 m wide with bedrock walls 40 m high and a gravel bed channel. During

flooding flow is hydraulically controlled by the narrow canyon entrance producing a

hydraulic ponding upstream of the constriction. This favours the accumulation of

flood deposit benches upstream of the gorge entrance. Palaeoflood sedimentation also

occurred along the canyon sides in small bedrock alcoves and in small tributary

gullies in the canyon.

3. Methodology

Palaeoflood hydrology using slackwater flood deposits as stage indicators of floods

(Baker & Kochel 1988, Baker et al. 2002, Benito et al. 2003a) were used to complete

a hydrological record of extreme events. Stratigraphic and sedimentological analyses

of the deposits were carried out both in the field and the laboratory, with sediment

peels of the stratigraphic profiles, measuring approximately 80 cm x 50 cm in size,

made in the field (Hattingh & Zawada,1996; Thorndycraft et al., 2005). Individual

flood units were identified through a variety of sedimentological indicators (Baker &

Kochel 1988, Benito et al. 2003b): the identification of clay layers at the top of a unit;

erosion surfaces; bioturbation indicating the exposure of a sedimentary surface;

angular clast layers, where local alcove or slope materials were deposited between

flood events; and changes in sediment colour. As well as identifying individual flood

units, sedimentary flow structures were also described in order to elucidate any

changing dynamics during a particular flood event and/or infer flow velocities that

can improve discharge estimation (Benito et al. 2003c).

Slackwater flood deposit chronology was determined using radiocarbon dating of

charcoal collected from individual flood units. Necessary preparation and pre-

treatment of the sample material for radiocarbon dating was carried out by the 14C

laboratory of the Department of Geography at the University of Zurich (GIUZ). The

dating itself was done by AMS (accelerator mass spectrometry) with the tandem

accelerator of the Institute of Particle Physics at the Swiss Federal Institute of

Technology, Zurich (ETH). Calibration of the radiocarbon dates was carried out using

the CalibeETH 1.5b (1991) programme of the Institute for Intermediate Energy

Physics ETH Zürich, Switzerland, using the calibration curves of Kromer & Becker

(1993), Linnick et al. (1986) and Stuiver & Pearson (1993). A summary of the

samples submitted for dating is presented in Table 1.

Documentary flood data was compiled to complement the palaeoflood data. Archival

flood evidence provides direct (e.g. flood date, duration, stage or spatial extent) or

indirect (e.g. post-flood damage repair) information of individual events, allowing

flood chronologies and socio-economic impacts to be compiled (Barriendos and

Coeur, 2004; Mudelsee et al. 2004; Brazdil et al. 2006; Thorndycraft et al., 2006). In

this study, the direct archival sources consulted included: 1) the Municipal Historical

Archive of Lorca (AHL); 2) the Municipal Historical Archive of Murcia; and 3) the

Cathedral Chapter Archive of Murcia. Further bibliographic sources used included

scientific and technical reports, local history works and non-systematic compilations

by historians (Hernández Amores, 1885; Couchoud, 1965; Diaz Cassou, 1977, 1993;

Fontana Tarrats, 1978; Torres and Calvo, 1975; Merino Alvarez, 1978; Barriendos

and Rodrigo, 2006).

The available information from the Municipal Historical Archive of Lorca dates from

the 1500s so any earlier data must be considered anecdotal in nature and probably

incomplete. At Lorca, the Guadalentín River is incised a few meters into its

floodplain, and the active channel forms a mixed anabranching and braided pattern

(according to the classification of Nanson and Knighton, 1996) with an ephemeral

channel of 100 m width and a gradient of 0.0051. The channel bed mobility of the

Guadalentín river at Lorca, as well as the history of urban change, prevents a precise

estimation of historical flood discharges. A classification of historical flood

dimensions was carried out according to the three categories proposed by Barriendos

and Coeur (2004), namely ordinary, extraordinary and catastrophic flooding, where

qualitative flood damage and descriptions of the severity of overbank flow are

combined. In ordinary flood situations water discharge is contained within the channel

and banks, with no damages. An extraordinary flood resulted in localised overbank

flow, resulting in damage but without major destruction. The classification of a

catastrophic flood depends on flood water inundating large floodplain areas resulting

in general damage and destruction of infrastructure.

Discharge estimates were calculated by hydraulic modelling of the 1800 m surveyed

study reach using HEC-RAS (Hydrologic Engineering Center 1995). The computation

procedure is based on the solution of the one-dimensional energy equation, derived

from the Bernoulli equation, for steady gradually varied flow. Subcritical flow

conditions were assumed along the reach, with critical flow selected as the boundary

condition at the most downstream cross-section where the channel narrows to pass

through a bedrock section where a deep pool has been scoured. Manning’s n values of

0.03 and 0.04 for the valley floor and margins, respectively, were assigned. A

sensitivity test performed on the model shows that for a 25% variation in the

roughness values an error of 4% was introduced into the discharge results.

Palaeoflood discharge estimation was based on the calculation of the step-backwater

profile that best fitted the surveyed geomorphological evidence of flood stage at

different sites along the longitudinal profile (O´Connor and Webb, 1988). Rating

curves relating individual flood unit elevations with flood discharges were established

at each site. High water marks (e.g. drift wood and pockets of fluvial sand) from the

largest instrumental flood that occurred in 1973 were also fitted to the model to

estimate flood discharges of this event. The accuracy of the discharge estimates

depends on the stability of cross-section topography through time. At the study reach

channel aggradation has infilled the bedrock channel with gravel deposits susceptible

to scour and fill. Assuming this channel bed aggradation post-dates construction of

the Valdeinfierno reservoir (3 km downstream of the study reach), the calculated

discharges of earlier floods are likely to underestimate the real peak flow.

4. Results

4.1. Slackwater stratigraphy and sedimentology

Nine sites of slackwater flood deposition were described throughout the 2.5 km study

reach (Fig. 2). The five thickest flood deposit sites (ES1, ES2, ES3, ES4 and RM)

form benches in backwater areas upstream of the abrupt canyon constriction, where

water is temporarily back-flooded during large floods (Fig. 2). Pockets of flood

sediments are common throughout this reach and occur in high crevices (RML),

shelves (RCS), slope depressions and upstream of tributary gullies (TL1, TL2).

A composite site of two trenches ES1 and ES2 shows the most complete palaeoflood

record. The site is composed of a flood bench of up to 7 m in thickness on the right

margin of the Caramel River, 150 m upstream of the gorge entrance (Fig. 2). The

stratigraphic sections contain spectacular sequences of multiple fine-grained flood

deposits (Fig. 3), providing evidence of at least 24 individual flood layers deposited

over the last 1000 years (Fig. 4). The lower ten flood units were dated to between cal.

AD 890-1160 and 1000-1210 suggesting a period of frequent relatively small floods,

on the basis of the very fine and fine sand grain size and the thin stratigraphic layers.

The top of this flood sequence is overlain by slope deposits which indicate a break in

flood sedimentation of at least 250 years. Overlying this are three flood layers (the

lower one dated to cal. AD 1450-1650), which also have a fine to very fine sand

grain-size (Fig. 4). These flood layers are capped by a 42 cm-thick colluvial layer

indicating a second break in flood sedimentation. The upper part of the sequence is

represented by nine modern flood deposits, with a basal date of AD 1630-1890

(81.5%), a middle date of cal. AD 1815 ± 80, and the upper flood layer most probably

left by the 1973 flood (Fig. 4). Field evidence suggests that the 1973 flood was the

largest event over the last 1000 years. Fresh-looking sand deposits most probably

related to this flood are approximately 3 m higher than the uppermost flood layer at

the ES2 section.

The sedimentary sequences described at sites ES1 and ES2 can be interpreted in terms

of flow energy and suspended load sediment conditions (see Benito et al., 2003c). The

sedimentological characteristics of the oldest deposits, with finer grain sizes and

thinner beds than the upper (younger) deposits, may be indicative of smaller floods,

lower sediment loads or a different channel position. The evidence for small-to-

moderate magnitude flooding is also supported by the two dominant sedimentary

sequences at the lower section comprising: (1) massive and highly bioturbated fine to

very fine-grained sand, with occasional granules, capped by a massive bioturbated silt

with mud cracks; and (2) parallel lamination, sometimes resting on an erosional

surface that includes granules or mud clasts, with ripples located either in the middle

of the sequence or more often towards the top of the bed. By contrast the upper flood

units of ES2 show a coarser grain size and flow structures indicative of higher energy

conditions. In particular, units 17, 21 and 22 all exhibit an erosive surface overlain by

coarse to very coarse-grained sand with parallel lamination and three dimensional

bedforms (Fig. 5). These sometimes resemble hummocky cross-stratification, and

occasionally include clay chips. The upper part of these units are characterised by a

fine-grained highly bioturbated sand. This sedimentary sequence is interpreted as

being generated by extreme floods with high stream flow velocity (>1ms-1) even in

these ineffective flow areas (Fig. 5). Units 23 and 24 contain parallel lamination and

climbing ripples in phase to/or climbing-in-drift developed with medium or coarse-

grained sand with parallel lamination, overlain by medium to fine-grained sand with

climbing ripples (Fig. 4). Climbing ripples, in-drift or in-phase, result from a vertical

variation in velocity and sediment fallout rate from flow with a high sediment load.

High velocity and lower aggradation rates produce climbing-in-drift whereas lower

velocities and higher aggradation rates produce climbing-in-phase (Ashey et al., 1982;

Jopling and Walker, 1968). The two breaks in flood deposition, during which

colluvial deposition occurred, are indicative of little, if any, flood deposition and

were, therefore, interpreted as evidence of a lack of extreme events during those

periods.

At the Rambla Mayor (RM) site, around 200 m upstream of the gorge entrance, flood

deposits accumulated in a large bench throughout the left margin of the river channel

(Fig. 2). Site RM is a 3 m-thick section preserving at least twelve flood units (Fig. 4).

A basal unit at this site yielded a date of cal. AD 1790 ± 90, a middle unit was dated

to cal. AD 1820 ± 80 and the upper deposits, separated from the previous ones by

slope wash sediments, were radiocarbon dated as modern (Fig. 4). The stratigraphy of

this site is equivalent in the number of preserved layers and flood chronology to the

upper section of site ES2, with seven flood beds (units 1 to 7) likely to be equivalent

to ES2 units 16 to 20, and five flood beds (units 8 to 12) probably corresponding to

ES2 units 21 to 24. Note (Fig. 4) that site RM is 5 m lower in elevation than ES2 and

a higher number of flood beds is expected for the same time interval. Although it is

difficult to correlate individual palaeoflood units between RM and ES2, the

sedimentary structures at units 5 and 6 indicate very high energy conditions during

deposition. The sedimentology is characterised by medium to coarse-grained sand

with trough cross bedding (3D), and occasionally small planar cross bedding (2D).

These units contain parallel lamination produced by aggradation on a plane bed and

avalanching faces developed infilling previous small hollows or topographic

depressions. This sequence is interpreted as vertical accretion produced by 3D dune

migration in response to an increase in stream power. Vertical aggradation of fine

sand produced by plane bed migration occurs under flow conditions of either

increasing velocity or decreasing flow depth. Commonly, fluid escape structures or

convolute bedding are developed within the sequence by density inversion. At the top,

sub-aerial exposure is indicated by bioturbation. In the upper section of the RM site,

the highest energy conditions were reflected in the sedimentary sequence of unit 10

(Fig. 4). This coarsening upwards sequence is characterised by fine-grained sand with

climbing ripples in phase to/or climbing-in-drift, overlain by medium to coarse-

grained sand with parallel lamination (18 cm in thickness). This unit, therefore, may

relate to a large flood such as the 1973 event.

Other evidence for the 1973 flood is found in high-elevation sediment pockets

deposited on the valley slope at RML next to the RM profile (Fig. 2). The narrow

canyon carved in limestone is devoid of relevant flood deposits probably due to high

energy conditions associated with extreme flood events. Only modern deposits were

found in pockets on small bedrock shelves (RCS in Figs. 2 and 4) and within tributary

gullies of the canyon (TL1 and TL2; Figs. 2 and 4).

4.2. Palaeoflood hydraulics and discharge estimation At the ES2 stratigraphic profile (ES2), the lower 10 flood units (cal. AD 890-1160

and 1000-1210) are associated with estimated discharges of 15-580 m3s-1. The second

set of three flood units (cal. AD 1450-1650) is associated with minimum discharges of

580-680 m3s-1. Finally at ES2 the upper nine flood units post-dating cal. AD 1630-

1890 (81.5% probability range) matched discharges of 730-1000 m3s-1. The upper

most unit of the RM stratigraphic profile is associated with a minimum discharge of

500 m3s-1. These discharges are minimum estimates since the water depth above the

flood units is unknown, and therefore, discharge values may be considered as

conservative. High water marks of the 1973 flood found along the study reach

provided a discharge calculation of 1615 m3s-1, which shows that “real” peak

discharges could be as much as 35% higher than the minimum estimates.

4.3. Documentary flood evidence

During the Medieval period, severe flooding is reported to have occurred in the

Guadalentín at Murcia in AD1143 (Santa Lucia flood), as well as during the 13th (2

events), 14th (3 events) and 15 th (10 events) centuries. Since the 1500s the Municipal

Historical Archive of Lorca (MHAL) provides a more complete documentary record

of flood events and damages, although the objectivity or continuity of the information

is not guaranteed. During the period AD1500-1900, 31 flood events were reported in

the Municipal Archive, of which 15 were classified as ordinary, 7 extraordinary and 9

catastrophic (Fig. 6; Table 2). In addition, two floods resulted from dam failures at

Puentes in 1648 and 1802 (Bautista and Muñoz, 1986). Since the third Puentes dam

was completed on October 2nd 1884, flood occurrence in Lorca decreased. As a result,

most of the large floods since AD 1884 were reported within the ordinary flood

category at Lorca, although the peak discharges recorded at the Puentes reservoir are

indicative of a large magnitude in the upper catchment (Table 2).

The temporal distribution of historical floods shows four major flood clusters at AD

1648-1672, 1769-1802, 1830-1840 and 1877-1900, with an anomalous number of

catastrophic and extraordinary category floods during the 19th Century (Fig. 6).

According to the archive descriptions the most severe floods occurred in AD 1651,

1653, 1879 and 1973 accounting for more than 2600 fatalities in the Guadalentín-

Segura catchment, including the AD 1802 dam failure (Table 2).

5. Discussion 5.1. Combining palaeoflood and documentary flood records

The value of the correlation of the palaeo- and historical flood data from the

Guadalentín basin is to test the degree of long-term agreement between these two

types of flood archive in terms of the number of floods (frequency) and relative

magnitude (energy and/or discharge). Four assumptions need to be made for such a

correlation: (1) flood benches were built up by successive flood layer deposition that

increased the flood elevation threshold required for further sediment accumulation

(self-censoring cf. House et al., 2002); (2) all floods exceeding the censoring elevation

of the previous flood deposits left preserved sedimentary evidence; 3) documentary

records have maintained the same flood perception levels through time; and 4)

relative flood magnitudes at Lorca and at the palaeoflood study reach are similar.

Assumptions (1) and (2) imply that documentary floods classified as ordinary may not

be represented in the palaeoflood record except during the first stages of sediment

deposition (the events at the base of the stratigraphic profile). At the ES2 site these

floods were dated to AD 900-1200, prior to the documentary flood record. Since AD

1500 it is likely that most palaeofloods will correspond to either catastrophic or

extraordinary historical floods recorded at Lorca (Table 2). Assumption (4) may not

always be true due to the differences in catchment size. A large event in the upper

basin (at the palaeoflood reach) may not always result in a large event at Lorca. This

is especially so after flood regulation by the Valdeinfierno and Puentes reservoirs that

would have modified flood discharges downstream at Lorca whilst the palaeoflood

reach maintained a natural flood regime.

The stratigraphic record at Estrecho 2, located at a protected low energy site, indicates

that the palaeoflood record here is likely to preserve evidence of all major floods over

the last 500 years and probably since AD 1000. The stratigraphy shows a cluster of at

least three palaeoflood events (units 13, 14, 15) post-dating cal. AD 1450-1650 and

capped by a slope deposit unit which can be interpreted as a period of reduced high

magnitude flooding (Fig. 4). The documentary record contains three catastrophic

floods within the palaeoflood radiocarbon age range (AD 1568, 1651 and 1653),

followed by a period of 50 years without major (catastrophic and/or extraordinary)

flood events (Table 2; Fig. 6). The second half of the 17th Century witnessed at least

five ordinary floods which were associated with a period characterised by a negative

precipitation anomaly in southern Spain according to the rainfall reconstruction

provided by Rodrigo et al (1999). This period coincides in time with the accumulation

of slopewash deposits above flood unit 15 at ES2 site, and incipient soil development

and carbonate precipitation. During the first half of the 18th Century, the total number

of documented floods decreased in relation to the previous period, although three

large floods were reported at Lorca (AD 1704, 1728 and 1733) (Fig. 6; Table 2). The

last three decades of the 18th Century show an increase of the frequency of ordinary

floods (7 events) associated with an above average precipitation anomaly. The

palaeoflood record of the Rambla Mayor (RM) section shows a similar number of

floods in the lower sedimentary sequence, units 3 and 5 dated to 1795 ± 85 cal AD

and 1790 ± 90 cal AD, respectively. Only one flood unit was deposited at the higher

elevation in the stratigraphic section ES2 (unit 16), dated with a calibrated age of AD

1775 ± 95. Interestingly, unit 16 at ES2 contains a high degree of bioturbation

indicating a period of surface exposure after deposition. We speculate, therefore, that

unit 16 may correspond to the catastrophic 1733 flood at Lorca (Table 2), the

threshold censoring level precluding deposition at this site of the extraordinary floods

in AD 1704 and 1728, and the later 18th Century ordinary floods.

Both documentary and palaeoflood records show an increase in frequency of

extraordinary and catastrophic floods during the 19th Century, in particular during the

second half of the century (Table 2; Fig. 6). The severity of these floods is reflected in

both the documentary record, with high economic losses and casualties (e.g. the 1838

and 1879 events), and the palaeoflood record where high energy sedimentary

structures (e.g. hummocky type cross-bedding) are evident (Fig. 5). Interestingly, this

is the first time in the palaeoflood literature that hummocky type cross bedding has

been described in sedimentary environments related to slackwater flood deposition.

The radiocarbon age resolution for the last 250 years unfortunately prevents a robust

correlation between the documentary and palaeoflood records. Therefore, palaeoflood

units at ES2 were assigned to documentary flood years on the basis of the correlative

sequenced number of floods (catastrophic and extraordinary), together with minor

adjustments to match the most extreme documentary floods with palaeoflood units

containing high-energy sedimentary structures (units 17, 18, 21 and 22). The most

severe documented floods in the Guadalentín occurred in 1838 and 1879 and were

assigned to units 18 and 21 respectively (Table 2). In between these floods, the

stratigraphy shows at least two flood units (units 19 and 20; the former <7 cm in

thickness), whereas the documentary records contains only one catastrophic flood

(1860). An additional documented flood in AD 1877 classified in Lorca as ordinary

was assigned to unit 20.

Since 1884 the documentary flood record was influenced by the construction of the

third Puentes dam, reducing flood magnitude and altering flood perception levels at

Lorca. However, the gauge station data at the Puentes dam enabled comparison

between the palaeoflood and systematic records. The largest flood at the end of the

19th Century occurred in 1891, with a peak discharge of 1890 m3s-1 at Puentes (Table

2). This flood was assigned to unit 22, associated with an estimated minimum

discharge of 940 m3s-1. The second largest flood in the 20th Century occurred in 1941

(palaeoflood unit 23) and the largest in 1973 (unit 24). This flood recorded a peak

discharge of 3544 m3s-1 at Puentes (1428 km2), whereas the Valdeinfierno gauge

station (456 km2) recorded only 108 m3s-1 (daily discharge). The palaeoflood evidence

provided a minimum discharge of 1035 m3s-1 but the ubiquitous high water marks and

sand pockets along the study reach indicated a flow of 1615 m3s-1.

5.2. Environmental and climatic changes in the upper Guadalentín basin

The combined palaeoflood (P) and documentary (D) records indicate that past floods

were clustered during particular time periods, namely at: AD 950-1200 (≥10 P

events); 1648-1672 (≥2 P & ≥8 D events); 1769-1802 (≥2 P & ≥7 D events); 1830-

1840 (≥2 P & ≥4 D events); and 1877-1900 (≥3 P & ≥7 D events). These flood

clusters may be explained by climate change altering extreme rainfall patterns

reaching the basin and/or land-use change affecting runoff generation conditions. The

effect of climatic variability on flood hydrology may be reflected by changing runoff

volumes, peak discharge and frequency (Changnon and Demissie, 1996). Land-use

changes may be reflected not only in runoff generation but also in terms of sediment

delivery to stream channels (Trimble, 1983). It is expected that the most dramatic

land-use changes in the basin are related to human activities during historical times.

Environmental history may facilitate identification of the onset of severe human

impact, after which environmental change should be considered in addition to climatic

drivers modifying flood hydrology.

Due to the small size of the study basin it is difficult to obtain palaeoenvironmental

data specific to the catchment. However, valuable historical information on human

population and land-use changes since the Middle Ages is available for the Almeria

province in general (Sánchez-Picón et al., 1996; García Latorre et al., 1998, 2001),

and for the Almeria mountains in particular (Lentisco, 1996). The economic system of

the Moorish settlements in Almeria (AD 711-1570) was based on irrigation

agriculture centred on floodplains. Marginal areas and forests were left for hunting,

gathering and grazing, explaining the existence of a rich forest and fauna of the

Almeria mountains as described by the traveller J. Münzer in AD 1494 (Münzer,

1495). This type of agriculture required much less land to be farmed than cereal

agriculture introduced later on. According to Lentisco (1996) and García Latorre et

al., (1998) the upper Guadalentin catchment (1141 km2) was sparsely populated with

around 2184 inhabitants in AD 1573 (2 inhabitants per km2). After expulsion of the

Moors in the 16th Century, the population rose to 10,459 inhabitants (9 inhabitants per

km2) in AD 1753 (Fig. 6), and a shift in agriculture to farming of cereal crops

occurred (representing 16% of total land-use at the Almería province according to

García Latorre et al., 1998). By AD 1860 population had increased to 22,890 (20 per

km2), with 25-30% of land-use under cultivation causing environmental

transformation of the study region (Sanchez-Picón et al., 1996; Lentisco, 1996). The

accelerated environmental change included massive deforestation of millions of trees

of a variety of species (e.g. Pinus nigra, Pinus halepensis, Pinus pinaster and

Quercus ilex) to clear areas for the cultivation of marginal areas, as well as for

charcoal production and domestic use of firewood (Madoz, 1845; Lentisco, 1996;

García Latorre et al., 2001). Given this historical information it is most likely that

widespread deforestation occurred in the upper Guadalentín basin in the late 18th and

early 19th centuries. It is interesting to note that the sedimentological characteristics of

the palaeoflood deposits dated to the 19th Century (units 17-22) illustrate a significant

change in the hydrological and environmental conditions of the upper Guadalentín

catchment. The sedimentary structures indicate very high energy conditions during

flooding (e.g. hummocky type cross-bedding), whilst the anomalous increase in grain-

size and unit thickness (>40 cm) of the slackwater flood deposits (Fig. 4) also suggest

increased sediment loads in response to human impact in the basin. The role of

climate still needs to be highlighted, however, as this period also covers the last phase

of cold conditions (at AD 1810-1850 after Flohn, 1993) during the Little Ice Age

(LIA), and the transition towards the warmer conditions of the 20th Century. Figure 6

shows that flood clusters (Phases 4 and 5) during the 19th Century occurred within a

range of precipitation anomalies similar to that experienced in previous periods of 17th

Century (Phase 2) and late 18th Century (Phase 3). In contrast, however, the number

of extraordinary and catastrophic floods over the 19th Century was remarkably high.

In earlier periods, climatic variability may be considered as the major controlling

factor on the recorded changes in flood magnitude and frequency. During the last

millennium the first palaeoflood period (AD 950-1200) broadly corresponds to the

late Medieval Warm Period (probably AD 900-1200, after Flohn, 1993) though some

authors doubt that this period can be clearly defined at the global and continental scale

(Hughes and Diaz, 1994). The concentration of floods at 1050-750 cal. BP (AD 950-

1200) is also apparent in palaeoflood and documentary evidence from Atlantic basins

(see Fig. 1) of the Iberian Peninsula (Benito et al., 1996), these events probably

associated with unusually wet winters. Flooding of the Tagus River from AD 1100 to

1200 was frequent as well as exceptional, and water stage data indicate that these

floods were the greatest of the available documentary record (Benito et al., 2003b). Of

particularly large magnitude were the floods of AD 1168 (also recorded for the rivers

Duero and Guadalquivir; Benito et al., 1996), 1178, 1181 and 1207 (Benito et al.,

2003b).

The following period 750-550 cal. BP (AD 1200-1500) was characterised by a lack of

slackwater deposition at profiles ES1, ES2 and ES3, and accumulation of slope

deposits (with root marks, bioturbation and carbonate precipitation) at ES2. This

period, therefore, represents a phase of reduced flood frequency, also evident in the

Spanish palaeoflood record (Thorndycraft and Benito, 2006a, 2006b) and the

documentary flood record (Benito et al., 2003b). This contrasts to the alluvial

stratigraphy of the lower terrace of the Guadalentín near Librilla (downstream of

Lorca) where initial aggradation was dated to ca. 730-540 cal. BP (Calmel Avila 2000

and 2002). This period coincides with the main recent phases of floodplain

aggradation of Iberian rivers (Vita-Finzi, 1976; Butzer et al., 1985, Thorndycraft and

Benito, 2006b; Uribelarrea and Benito 2008). This observation may reflect the fact

that the two depositional environments (slackwater deposits and alluvial floodplains)

represent different event magnitude and frequency relationships (Thorndycraft and

Benito 2006a,b). The palaeoflood record preserves evidence of high magnitude, low

frequency floods. Alluviation, however, may result from a range of event magnitudes

that cause overbank flooding and/or lateral aggradation, as well as being a response to

an increase in sediment supply that can be brought about by human activity, as shown

by van Andel et al. (1990) in Greece and the Aegean.

The second and third flood clusters (1648-1672 and 1769-1802) coincide with the

LIA (ca 1500-1850), generally believed to be a widespread event at the global scale.

In the palaeoflood record, the AD 1200-1500 hiatus was broken by a sequence of

three palaeoflood events (units 13, 14 and 15 at ES2) post-dating AD1505-1636, and

assigned to historical floods at AD 1568, 1648 and 1651. Documentary evidence from

other Mediterranean river basins shows an increase of severe flooding during AD

1580-1620, particularly in Catalonian rivers in NE Spain (Barriendos and Martín

Vide, 1998; Llasat et al., 2005). This period overlaps with one of the coldest phases

(AD 1590-1650) of the Little Ice Age, the Maunder Minimum, during which sunspots

were exceedingly rare (Vaquero et al., 2002). In southern Spain prevailing anomalies

were wet, with the wettest years centred on the decades of the 1590s, 1630s and 1640s

(Barriendos, 1997, Rodrigo et al., 1999 and Fig. 6), with large agricultural losses due

to excess rainfall noted in Murcia Province (Barriendos, 1997).

The flood cluster at AD 1648-1672 (2) is mainly composed of ordinary floods (with

the exception of the 1648 and 1651 events) and no links to anomalous behaviour of

regional climate conditions have been found. Similarly, seven documentary flood

events were described between AD 1769-1802 (3), overlapping the period AD 1760-

1800, which was noteworthy for its strong climatic variability in the Mediterranean

region with both floods and droughts occurring (Barriendos and Llasat, 2003). It can

be noted that flood generating conditions, in some decades of period 3, may be

enhanced under positive precipitation anomalies at the regional scale, although few

extraordinary and catastrophic floods were recorded from the 1500s to 1830s. In the

19th Century the anomalously high number of extraordinary and catastrophic floods,

occurred even during decades of negative or close to normal precipitation anomalies,

suggest that the climate signal of flood response was amplified by human disturbance.

This severe human impact shows the high sensitivity of Mediterranean basins to soil

and hydrological disturbance, which in the space of a few decades led to a high degree

of degradation in the upper Guadalentín catchment.

6. Conclusions

A combined palaeo and historical flood record provides evidence of extreme event

response to global change over the last 1000 years in the Guadalentín basin (SE

Spain). The stratigraphic and documentary records identify five main phases of

increased flood frequency. Phase 1, on the basis of sedimentary palaeoflood evidence

only, occurred ca. 950-1200 cal AD with at least ten floods with minimum discharge

estimates of 15-580 m3s-1. Phases 2-5 (Fig. 6), based on combined sedimentary and

documentary evidence, occurred at: (a) AD 1648-1672, recording eight documentary

floods and two palaeofloods exceeding 580-680 m3s-1; (b) AD 1769-1802, comprising

seven documentary floods, probably two included in the sedimentary record (at the

RM site) with minimum discharges of 250 m3s-1; (c) AD 1830-1840 recording four

documentary floods, with at least two events exceeding a discharge of 760-1035 m3s-1

recorded in the stratigraphy; (d) AD 1877-1900 includes seven documentary floods,

three of those exceeding 880 m3s-1 provided sedimentary evidence. In addition, two

flood units represented in the upper sequence of the ES2 stratigraphic profile are

probably related to the AD 1941 and 1973 floods, the latter being the largest flood in

the gauge record at Puentes. Discharge estimation at the palaeoflood study reach was

complicated by gravel deposition within the bedrock channel and subsequent

problems of scour and fill. As a result it is likely that for the older flood deposits the

calculated values are probably underestimates of the real discharges. Nevertheless, the

excellent flood frequency record at the site, in combination with the documentary

flood evidence, allows discussion of the roles of climatic variability and human

impact in driving the changes in flood frequency over the last 1000 years. Phase 1

coincides with the MWP, although detailed palaeoclimate reconstructions are limited

for this time period so it is not known whether floods relate specifically to a warming,

warm or cooling phase of climate. Greater precision can be attributed to phases 2-5,

which occurred during the LIA where both the flood record and climatic history is

better understood through corroboration between palaeoenvironmental and

documentary evidence. Phases 2 and 3 occurred during known cold periods of the

LIA, with wet anomalies (phase 3) in southern Spain, and correlate with flooding

elsewhere in Spain (Llasat et al., 2005), including large Atlantic basins such as the

Tagus (Benito et al., 2003a, 2003b). Phases 4 and 5 represent the most severe flooding

in terms of high energy conditions evident in the sedimentary structures as well as the

socio-economic impacts of the events at Lorca. This period not only represents the

last phase of the LIA and the change to 20th Century warmer conditions but also

coincides with historical evidence of major economic and land-use transitions in the

Almeria mountains, with high rates of population increase, major deforestation and an

increase in cultivated land area into marginal areas. Although the flooding cannot be

separated from climatic variability it appears that human impact has had a major

influence on the sensitivity of flood hydrology and the subsequent severity of the

events. This paper highlights many of the complexities associated with understanding

flood response to global change and shows the value of palaeo and historical flood

data for elucidating flood magnitude and frequency relationships under past global

change scenarios.

Acknowledgements

This research was funded by the EC through the projects “Systematic, Palaeoflood

and Historical data for the improvEment of flood Risk Estimation – SPHERE”

(contract no. EVG1-CT-1999-00010) and “FloodWater recharge of alluvial Aquifers

in Dryland Environments – WADE (contract no. GOCE-CT-2003-506680). The

Spanish Commission of Science and Technology grant HID99-0858, and the project

“Infiltration on channel beds and recharge of aquifers related to floods and

palaeofloods in ephemeral rivers- PALEOREC” (CICYT project no.CGL2005-

01977/HID). The authors are very grateful to Tim van der Schriek and Noam

Greenbaum for their critical review of the original manuscript and for their helpful

comments and suggestions.

References Ashley, G.M., Southard, J.B., Boothroyd, J.C., 1982. Deposition of climbing-ripple

beds: a flume simulation. Sedimentology 29, 67-79.

Baker, V.R., Kochel, R.C., 1988. Flood sedimentation in bedrock fluvial systems. In:

Baker, V.R., Kochel, R.C., Patton, P.C. (Eds.), Flood Geomorphology. John

Wiley & Sons Ltd., U.S.A., pp. 123-137.

Baker, V.R., Kochel, R.C., Patton, P.C., Pickup, G., 1983. Palaeohydrologic analysis

of Holocene flood slack-water sediments. Spec. Publs. Int. Ass. Sediment. 6,

229-239.

Baker, V.R., Webb, R.H., House, P.K., 2002. The Scientific and societal value of

paleoflood hydrology. In: House, P.K., Webb, R.H., Baker, V.R., Levish, D.R.

(Eds.), Ancient Floods, Modern Hazards: Principles and Applications of

Paleoflood Hydrology, Water Science and Application Series, Vol. 5, pp. 127-

146.

Barriendos, M., 1997. Climatic variations in the Iberian Peninsula during the late

Maunder Minimum (AD 1675-1715): an analysis of data from rogation

ceremonies. The Holocene 7, 1, 105-111.

Barriendos, M., Martín Vide, J., 1998. Secular Climatic Oscillations as Indicated by

Catastrophic Floods in the Spanish Mediterranean Coastal Area (14th-19th

Centuries). Climatic Change 38, 473-491.

Barriendos, M., Llasat, M.C., 2003. The Case of the `Maldá' Anomaly in the Western

Mediterranean Basin (AD 1760–1800): An Example of a Strong Climatic

Variability. Climatic Change 61, 191-216.

Barriendos, M., Coeur, D., 2004. Flood data reconstruction in historical times from

non-instrumental sources in Spain and France. In: Benito, G., Thorndycraft,

V.R. (Eds.), Systematic, Palaeoflood and Historical Data for the Improvement

of Flood Risk Estimation. Methodological Guidelines. Centro de Ciencias

Medioambientales, Madrid, Spain, pp. 29-42.

Barriendos, M., Rodrigo, F.S., 2006. Study on historical flood events of spanish rivers

using documentary data. Hydrological Sciences Journal 51, 5, 765-783.

Bautista, J., Muñoz, J., 1986. Las presas del estrecho de Puentes. Confederación

Hidrográfica del Segura, Murcia.

Benito, G., Machado, M.J., Pérez-González, A., 1996. Climate change and flood

sensitivity in Spain. In: Branson, J., Brown, A.G., Gregory, K.J. (Eds.), Global

continental changes: The context of palaeohydrology. Geological Society

Special Publication 115, London, pp. 95-98.

Benito, G., Sopeña, A., Sánchez-Moya, Y., Machado, M.J., Pérez-Gonzalez, A.,

2003a. Palaeoflood record of the Tagus River (Central Spain) during the Late

Pleistocene and Holocene. Quaternary Science Reviews 22, 1737-1756.

Benito, G., Diez-Herrero, A., de Villalta, M.F., 2003b. Magnitude and frequency of

flooding in the Tagus basin (Central Spain) over the last millennium. Climatic

Change 58, 171-192.

Benito, G., Sánchez-Moya, Y., Sopeña, A., 2003c. Sedimentology of high-stage flood

deposits of the Tagus River, Central Spain. Sedimentary Geology 157, 107-

132.

Benito, G., Lang, M., Barriendos, M., Llasat, M.C., Frances, F., Ouarda, T.,

Thorndycraft, V., Enzel, Y., Bardossy, A., Cœur, D., Bobée, B., 2004. Use of

Systematic Paleoflood and Historical Data for the Improvement of Flood Risk

Estimation. Review of Scientific Methods. Natural Hazards 31, 623-643.

Brázdil, R., Kundzewicz, Z.W., Benito, G., 2006. Historical hydrology for studying

flood risk in Europe. Hydrological Sciences Journal 51 (5), 739-764.

Butzer, K.W., 1974. Accelerated soil erosion. In: Manners, I., Mikesell, M.W. (Eds.),

Perspectives on Environment. Association of American Geographers,

Washington D.C., pp. 57-78.

Butzer, K.W., 1975. Pleistocene littoral-sedimentary cycles of the Mediterranean

Basin: a Mallorquin view. In: Butzer, K.W., Isaac, G.L. (Eds.), After the

Australopithecines. Mouton, The Hague, pp. 25-71.

Butzer, K.W., Butzer, E.K., Mateu, J.F., Kraus, P., 1985. Irrigation agrosystems in

Eastern Spain: Roman or Islamic origins?. Annals, Association of American

Geographers 75, 495-522.

Calmel-Avila, M., 2000. Procesos hídricos holocenos en el bajo Guadalentín (Murcia,

SE España). Cuaternario y Geomorfología 14, 65-78.

Calmel-Avila, M., 2002. The Librilla “rambla”, an example of morphogenetic crisis in

the Holocene (Murcia, Spain). Quaternary International 93-94, 101-108.

Capel, J., 1974. Génesis de las inundaciones de Octubre de 1973 en el Sureste de la

Península Ibérica. Cuadernos Geográficos 4, 149-166.

Capel, J., 1981. Los Climas de España. Col. Ciencias Geográficas. Oikos-Tau,

Barcelona.

Changnon, S.A., Demissie, M., 1996. Detection of changes in streamflow and floods

resulting from climate fluctuations and land use-drainage changes. Climatic

Change 32, 411-421.

Couchoud, R., 1965. Hidrología histórica del Segura. Efemérides hidrológica y

fervorosa, Centro de Estudios Hidrográficos, Madrid.

Diaz, P., 1977. Serie de los obispos de Cartagena. Sus hechos y su tiempo. Instituto

Municipal de Historia, Murcia.

Diaz, P., 1993. Topografía, geología, climatología de la Huerta de Murcia. Librerías

París-Valencia.

Ely, L.L., 1997. Response of extreme floods in the south-western United States to climatic variations in the late Holocene. Geomorphology 19, 175-201.

Flohn, H., 1993. Climatic Evolution During the Last Millennium: What Can We

Learn from It?. In: Eddy, J.A., Oeschger, H. (Eds.), Global Changes in the

Perspective of the Past. John Wiley and Sons Ltd., Chichester, pp. 295-316.

Fontana, J.M., 1978. Historia del Clima en el Litoral Mediterráneo: Reino de Valencia

mas Provincia de Murcia. Unpublished report, Jávea.

Foulds, S.A., Macklin, M.G., 2006. Holocene land-use change and its impact on river

basin dynamics in Great Britain and Ireland. Progress in Physical Geography

30, 589-604.

Frei, C., Schär, C., 2001. Detection of probability of trends in rare events: Theory and

application to heavy precipitation in the Alpine region. Journal of Climate 14,

1568–1584.

Garcia Latorre, J., 1998. La agricultura almeriense antes y después de la expulsión de

los moriscos. Chronica Nova 25, 275-300.

Garcia Latorre, J., Garcia-Latorre, J., Sanchez-Picon, A., 2001. Dealing with aridity:

socio-economic structures and environmental changes in an arid

Mediterranean region. Land Use Policy 18, 53-64.

Gregory, K. J., Benito, G., Dikau, R., Golosov, V., Jones A.J.J., Macklin M.G.,

Parsons, A,J., Passmore D.G., Poesen, J., Starkel, L., Walling, D.E., 2006.

Past hydrological events related to understanding global change: An ICSU

research project. Catena 66, 2-13.

Hattingh, J., Zawada, P.K., 1996. Relief peels in the study of palaeoflood slack-water

sediments. Geomorphology 16, 121-126.

Hernández, A., 1885. Inundaciones de la huerta de Murcia. Imprenta El Diario, Murcia.

Hirschboeck, K.K., 1991. Climate and floods. U.S. Geological Survey Water Supply

Paper 2375, 67-88.

House, P.K., Pearthree, P.A., Klawon, J.E., 2002. Historical flood and paleoflood

chronology of the lower Verde River, Arizona: stratigraphic evidence and

related uncertainties. In: House, P.K., Webb, R.H., Baker, V.R., Levish, D.R.

(Eds.), Ancient Floods, Modern Hazards: Principles and Applications of

Paleoflood Hydrology. American Geophysical Union, Water Science and

Application Series 5, pp. 267-293.

Hughes, K.M., Diaz, H.F., 1994. Was There a ‘Medieval Warm Period’, and if so,

Where and When?. In: Hughes, M.K., Diaz, H.F. (Eds.), The Medieval Warm

Period. Kluwer Academic Publishers, Dordrecht (The Netherlands), pp. 109-

142.

Hydrologic Engineering Center, 1995. HEC-RAS, River Analysis System, Hydraulics

Reference Manual, (CPD-69).

Jopling, A.L., Walker, R.G., 1968. Morphology and origin of ripple-drift cross-

lamination, with examples from the Pleistocene of Massachusetts. J. Sedim.

Petrol. 38, 971-984.

Klein Tank, A.M.G., Können, G.P., 2003. Trends in indices of daily temperature and

precipitation extremes in Europe, 1946–1999. Journal of Climate 16, 3665-

3680.

Knighton, A.D., 1998. Fluvial forms and processes: A new perspective. Arnold,

London.

Kromer, B., Becker, B., 1993. German oak and pine C-14 calibration, 7200-

9439 BC. Radiocarbon 35, 125-135.

Lewin, J., Macklin, M.G., Woodward, J.C., 1995. Mediterranean Quaternary River

Environments- Some future research needs. In: Lewin, J., Macklin, M.G.,

Woodward, J.C. (Eds.), Mediterranean Quaternary River Environments.

Balkema, Rotterdam, pp. 283-284.

Lentisco, J.D., 1996. De despensa agrícola a recurso turístico. Evolución y

transformación del Monto Velezano (S. XVI-XX). In: Sánchez-Picón, A.

(Ed.), Historia y medio ambiente en el territorio almeriense. Universidad de

Almería, Almería, pp. 203-220.

Linnick, T.W., Long, A., Damon, P.E., Ferguson, C.W., 1986. High-precision

radiocarbon dating of bristlecone-pine from 6554 to 5350 BC. Radiocarbon

28, 943-953.

Llasat, M.C., Barriendos, M., Barrera, A., Rigo, T., 2005. Floods in Catalonia (NE

Spain) since the 14th century. Climatological and meteorological aspects from

historical documentary sources and old instrumental records. Journal of

Hydrology 313, 32-47.

Madoz, P., 1845. Diccionario geográfico-estadístico-histórico de España y sus

posesiones de Ultramar, (1845-1850). Editoriales Andaluzas Unidas, Edition

of 1988. 16 vols.

Merino, A., 1978. Geografía histórica del territorio de la actual provincia de Murcia

desde la Reconquista, Academia Alfonso X el Sabio, Murcia. Edición facsímil

de 1915.

Mudelsee, M., Börngen, M., Tetzlaff, G., Grünewald, U., 2004. Extreme floods in

central Europe over the past 500 years: Role of cyclone pathway “Zugstrasse

Vb”. Journal of Geophysical Research 109, D23101.

Muñoz, J., 1989. Enseñanza de las avenidas históricas en la Cuenca del Segura. In:

Gil, A., Morales, A. (Eds.), Avenidas fluviales e inundaciones en la Cuenca

del Mediterráneo. Instituto Universitario de Geografía, Universidad de

Alicante, pp. 459-467.

Münzer, J., 1495. Viaje por España y Portugal (1494-1495). Colección El Espejo

Navegante. Ediciones Polifemo, 2nd edition, Madrid 2002.

Nanson, G.G., Knighton, A.D., 1996. Anabranching rivers: their cause, character and

classification. Earth Surface Processes and Landforms 21, 217-239.

O'Connor, J.E., Webb, R.H., 1988. Hydraulic modeling for paleoflood analysis. In:

Baker, V.R., Kochel, R.C., Patton, P.C. (Eds.), Flood Geomorphology. John

Wiley & Sons, United States, pp. 393-402.

Pardé, M., 1961. Sur la puissance des crues en diverses parties du monde.

Geographica, CSIC, Zaragoza.

Rodrigo, F.S., Esteban-Parra, M.J., Pozo-Vázquez, D., Castro-Díez, Y., 1999. A 500-

year precipitation record in southern Spain. Internacional Journal of

Climatology 19, 1233-1253.

Sánchez-Picón, A., 1996. La presión humana sobre el monte en Almería durante el

siglo XIX. In: Sánchez-Picón, A. (Ed.), Historia y medio ambiente en el

territorio almeriense. Universidad de Almería, Almería, pp. 169-220.

Sheffer, N.A., Rico, M., Enzel, Y., Benito, G., Grodek, T., 2007.The palaeoflood

record of the Gardon river, France: A comparison with the extreme 2002 flood

event. Geomorphology. doi: 10.1016/j.geomorph.2007.02.034

Stuiver, M., Pearson, G.W., 1993. High-precision bidecadal calibration of the

radiocarbon timescale, AD 1950-500 BC and 2500-6000 BC. Radiocarbon 35,

1-23.

Thorndycraft, V.R., Benito, G., 2006a. The Holocene fluvial chronology of Spain:

Evidence from a newly compiled radiocarbon database. Quaternary Science

Reviews 25, 223-234.

Thorndycraft, V.R., Benito, G., 2006b. Late Holocene fluvial chronology of Spain:

The role of climatic variability and human impact. Catena 66, 34-41.

Thorndycraft, V.R., Benito, G., Rico, M., Sopeña, A., Sánchez-Moya, Y., Casas-

Planes, A., 2004. A Late Holocene Paleoflood record from slackwater flood

deposits of the Llobregat River, NE Spain. Journal Geological Society of India

64(4), 549-559.

Thorndycraft, V.R., Benito, G., Rico, M., Sopeña, A., Sánchez, Y., Casas, M., 2005.

A long-term flood discharge record derived from slackwater flood deposits of

the Llobregat River, NE Spain. Journal of Hydrology 313 (1-2), 16-31.

Thorndycraft, V.R, Barriendos, M., Benito, G., Rico, M., Casas, A., 2006. The

catastrophic floods of AD 1617 in Catalonia (northeast Spain) and their

climatic control. Hydrological Sciences Journal 51 (5), 899-912.

Torres, J., Calvo, F., 1975. Inundaciones en Murcia (siglo XV). Papeles del

Departamento de Geografía, Universidad de Murcia, Murcia, 29-50.

Trenberth, K.E., Jones, P.D., Ambenje, P., Bojariu, R., Easterling, D., Klein Tank, A.,

Parker, D., Rahimzadeh, F., Renwick, J.A., Rusticucci, M., Soden, B., Zhai,

P., 2007. Observations: Surface and Atmospheric Climate Change. In:

Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B.,

Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science

Basis. Contribution of Working Group I to the Fourth Assessment Report of

the Intergovernmental Panel on Climate Change. Cambridge University Press,

Cambridge, United Kingdom and New York, NY, USA, pp. 235-336.

Trimble, S.W., 1983. A sediment budget for Coon Creek basin in the Driftless Area,

Wisconsin, 1853-1977. American Journal of Science 283, 454-474.

Trumbore, S.E., 2000. Radiocarbon geochronology. In: Noller, J.S., Sowers, J.M.,

Lettis, W.R. (Eds.), Quaternary Geochronology: Methods and Applications.

AGU, Washington, D.C., pp. 41– 60.

Uribelarrea, D., Benito, G., 2008. Fluvial changes of the Guadalquivir river during the

Holocene in Córdoba (S. Spain). Geomorphology, in press.

van Andel, J.H., Zangger, H., Demitrack, A., 1990. Land use and soil erosion in

prehistoric and historic Greece. Journal of Field Ecology 17, 379-396.

Vaquero, J.M., Sánchez-bajo, F., Gallego, M.C., 2002. A Measure of the Solar

Rotation During the Maunder Minimum. Solar Physics 207 (2), 219-222.

Vita-Finzi, F., 1969. The Mediterranean valleys: Geological changes in historical

times. Cambridge University Press, Cambridge.

Vita-Finzi, F., 1976. Diachronism in Old World alluvial sequences. Nature 263, 218-

219.

River Flood unit Sample material

Lab code Age (yrs BP)

Calibrated age

(years AD)

Two sigma calibrated age

range

Known historical flood(s)

ES-1-T4

Charcoal UZ-4597/ ETH-24410

980±45

1080±45

AD 980-1190

ES-2-T14

Charcoal UZ-4598/ ETH-24411

945±45

1105±50

AD 1000-1210 1143 Santa Lucía flood

ES-2-T17

Charcoal UZ-4599/ ETH-24412

1020±50

1035±60

AD 890-1160

ES-2-T21

Charcoal UZ-4659/ ETH-24681

190±55

1790±95

AD 1630-1960

ES-2-T23

Charcoal UZ-4600/ETH-24413

340±45

1560±55

AD 1450-1650

ES-2-T29

Charcoal UZ-4601/ETH-24414

205±45

1775±95

AD 1630-1890 (81.5%)

1830

ES-2-T31

Charcoal UZ-4602/ETH-24415

120±45

1815±80

AD1670-1960

ES-3-T6

Charcoal UZ-4603/ETH-24416

1985±50

30±55

110 BC – AD 130 47 BC Julius Caesar flood

Caramel

ES-3-T9

Charcoal UZ-4660/ETH-24682

120±55

1810±80

AD 1660-1960

RM1-3

Charcoal UZ-4994/ETH-27650

175±45

1795±85

AD 1650-1890

RM1-5

Charcoal UZ-4995/ETH-27651

190±45

1790±90

AD 1640-1960

RM1-6

Charcoal UZ-4996/ETH-27652

105±45

1820±80

AD 1670-1960

Rambla Mayor

RM1-11 Charcoal UZ-4997/ETH-27653 Modern Table 1. Radiocarbon dating samples and results, including calibrated ages calculated by the CalibETH 1.5b programme using the calibration curves of Kromer and Becker (1993), Linick et al. (1986) and Stuiver and Pearson (1993).

Year Date Relative Magnitude

Comments Flood unit in profile

Estrecho 2

Q (m3s-1) recorded in

lower catchment (Puentes)

Minimum Q (m3s-1) at upper

catchment (this study)

1568 17th September Catastrophic - 13 - 625 1648 5th August Ordinary Damage on Puentes Dam

foundation - - -

1651 14th October Catastrophic San Calixto flood 1000 deaths in Murcia region Destruction of > 1000 houses

14 - 640

1653 4th November Catastrophic San Severo flood In lowland areas of Murcia: >250 deaths in Murcia region Destruction of 4000 houses out of 6000

15 - 680

1704 27th August Extraordinary San Leovigildo - - - 1728 29th October Catastrophic - - - - 1733 3rd September Catastrophic Nª Sª de los Reyes flood 16 - 730 1802 30th April Ordinary 2nd Breakdown of Puentes

Dam; 608 deaths Catastrophic only downstream of the dam

- - -

1830 3rd September Extraordinary - - - - 1831 19th October Extraordinary La Diforme flood 17 - 760 1838 3rd October Catastrophic San Francisco flooding

Several deaths in Lorca 18

- 810

1860 17th September Catastrophic - 19 - 870 1877 19th September Extraordinary San Cosme and San

Damian- 20 - 880

1879 14th October Catastrophic Santa Teresa flood 761 deaths in Murcia; 13 in Lorca, 2 in Librilla and 1 in Cieza

21 1510 915

1880 28th August Catastrophic - - - - 1884 22nd May Catastrophic Ascensión flood - 1136 - 1891 11th September Ordinary* San Jacinto flood 22 1890 940 1898 16th January Extraordinary San Fulgencio flood - - - 1900 27th June Ordinary* San Aniceto flood - 1295 - 1921 24th September Ordinary* Virgen de las Mercedes

flood - 1005 -

1941 28th June Ordinary* - 23 1378 980 1973 19th October Catastrophic 300 mm in 24 h; 13 deaths

in Lorca and 83 in Puerto Lumbreras

24 3544 1035-1616

Table 2. Chronology of the major Guadalentín floods and assigned flood units of the ES 2 profile (number refers to stratigraphic profile in Fig. 4). Relative flood magnitude follows the classification by Barriendos and Coeur, 2004. Minimum discharges were estimated using sedimentary evidences as palaeostage indicator. *Flood hydrograph attenuated at Lorca by the Puentes reservoir.

Figure captions

Figure 1. Location of the Guadalentín river catchment and the study area.

Figure 2. Geomorphological sketch map of the study reach illustrating the location of slackwater flood deposits and the stratigraphic profiles. Figure 3. View of the ES 2 stratigraphic profile showing the spectacular sequence of slackwater flood deposition reaching a thickness of ca. 7 m. Flood unit numbers used in the text are annotated. Figure 4: Stratigraphic profiles at the study reach (radiocarbon dates in calibrated years AD) and proposed correlations between sections. Figure 5. A : Photograph of lacquer peel (50 cm high) of slackwater flood deposits

(units 16, 17 and 18) from the ES2 site; B: Grain size (VC: very coarse, C; coarse, M;

medium, F; fine, Vf; Very fine, L; Silt ); C: Flood sequence number (see ES2 profile,

Fig. 4); D: Sketch drawn from lacquer peel; and E: Inferred flood velocity

Figure 6. Upper: (i) Reconstruction of precipitation anomalies 1501–1990 for southern Spain (10 year moving average) after Rodrigo et al. (1999). The anomalies for AD 1500-1791 were estimated from documentary records, and since 1791 from instrumental data. (ii) Normalised annual flood series (this study) smoothed with a Gaussian filter (11 and 30 years, respectively). Lower: (a) Accumulated number of documentary floods compiled from the Municipal Archive of Lorca. The relative magnitudes of the events are indicated (classified according to Barriendos and Coeur, 2004). (b) Accumulated number of palaeoflood events. Palaeoflood units 17, 18, 21 and 22 (profile ES2) containing high-energy sedimentary structures are identified; other units of ES2 follow a correlative order. (c) Time framework of slope accumulation in the sedimentary record. (d) Demographic evolution (inhabitants/km2) of the upper Guadalentin region (1141 km2) according to Lentisco (1996). Shaded areas show the temporal distribution of flood clusters (Phases 2-5) described in the text. Phase 1, not shown, relates to the palaeoflood evidence of increased flood frequency dated to cal. AD950-1200.

Fig.1

Fig. 2

Fig. 3

24

23

21

22

20

18

17

16

15

13

14

11

70m

0.5

1

89

10

687,5 m

17

23

1615

1413

7

6

9

1

Jurassic

CO3=

CO3=

AD 1105 50

AD 1035 60

AD 1560 55

AD 1775 95

AD 1815 80

AD 30 55

692,2 m

684,99 m

1

2

3

56

79

8

Rambla

?

??

Rambla

Rambla

12

11

10

9

8

7

6

5

43

21

687,39 m

693,5 m

1

692,9 m

686,7 m

1

112

3

14C

Fresh sand, likely 1973 flood

14C

14C

14C

14C

14C

14C

14C

14C

14C

14C

14C= Modern

soil

8

76

5

4

2

3

1

680,85

LEGEND

Trough (3d)Cross bedding

Bioturbation

Ripples lamination

HummokyCross bedding

Roots

Climbing ripples

Parallel lamination

Climbing ripplesIn phase

Planar (2d)Cross bedding

Charcoal14C

Slope washflow

CarbonatesC03

=

Bedrock

Bedrock

Bedrock

Bedrock

BedrockBedrock

Gravel levels

Mud cracks

1 2

2

Fresh sand, likely 1973 floodFresh sand, likely 1973 flood

AD 1080 45

C+sl = Clay-Silt

S = NUMBER OF SEQUENCE( )each represents one flood only

Clay

C+sl Sand Gravel

vf f m c vc

GRAIN SIZE

AD 1080 45 = RADIOCARBON CALIBRATED AGE

Soil

RAMBLA CARAMELSHELTER (RCS)

ESTRECHO 2 (ES2)

ESTRECHO 1 (ES1)

ESTRECHO 4 (ES4)

ESTRECHO 3 (ES3)

TRIBUTARY L2 (TL2)TRIBUTARY L1 (TL1)

RAMBLA MAYOR (RM)

RAMBLA MAYOR L (RML)

AD 1810 80

AD 1795 85

AD 1790 90

AD 1820 80

Fig. 5

Fig. 6