Historical flood data series of Eastern Spanish Coast (14th-20 th centuries). Improving identification of climatic patterns and human factors of flood events 1 2 3 from primary documentary sources Armando Alberola (1), Mariano Barriendos (2), Salvador Gil-Guirado (3), Alfredo Pérez-Morales (3), Carles Balasch (4), Xavier Castelltort (4), Jordi Mazón (5), David Pino (5), Josep Lluís Ruiz-Bellet (4), and Jordi Tuset (6) (1) Area de Historia Moderna, Universidad de Alicante, Alicante, , (2) Departament d’Història Moderna, Universitat de , Barcelona, Spain, (3) Departamento de Geografía, Universidad de Murcia, Murcia, Spain, (4) Departament de Medi Ambient i Ciències del Sòl, Universitat de Lleida, Lleida, Spain, (5) Departament de Física Aplicada, Universitat Politècnica de Catalunya, Barcelona, Spain, (6) Centre Tecnològic Forestal de Catalunya, Solsona, Spain 4 5 6 A. Introduction B. Methodology Historical flood events in eastern Spanish coast have been studied by different research groups and projects. Complexity of flood processes, involving Along the entire east of the Iberian Peninsula, we have 15 series of The collected floods require a minimal common characterization in order to be classified. Most of the flood records are still to be completed with more precise and reliable information atmospheric, surface and human factors, is not easily understandable when long time series are required. Study area floods with at least four centuries extension (12 in , search, but, for the moment, the most evident traits can be used. The more common elements to an event of any time are those referring to its basic hydrological behaviour and the Present analysis from PREDIFLOOD Project Consortium defines a new step of flood event databases: Improved access to primary (documentary) and secondary 1 and 2 in the Region of Murcia). impacts it caused. The combination of these two criteria has been used in this study to create the following classification system by assessment of impacts: (bibliographical) sources, data collection for all possible locations where floods are detected, and improved system of classification (Barriendos et al., 2014). The eastern Spanish coast is made up of three regions, which is 1. Non-overbank flood + disturbance + damage: Extraordinary flood A first analysis is applied to 8 selected flood series. Long chronologies from PREDIFLOOD Project for Catalonia region (Girona, Barcelona, Tarragona, Lleida, composed in 8 provinces. The surface of this area is 66,682 square 2. Overbank flood + disturbance: Extraordinary flood Tortosa). In addition, to cover all sector of spanish mediterranean coast, we introduce Valencia city in Turia River basin. South Eastern sector is cover with kilometers (13.2% of of Spain) and the population is 13.99 million 3. Overbank flood + disturbance + damage: Extraordinary/catastrophic flood Murcia and Caravaca cities, Segura River basin. (30.1% of Spain). 4. Overbank flood + damage + destruction: Catastrophic flood Extension of area under study required contributions of research teams experienced in work of documentary primary sources (Alberola, 2006; Gil-Guirado, In this area much of the dynamism, population and economic In general, the basic criteria are the occurrence of flood and whether it is an overbank flood or not. Then, there are two further levels: first, the capacity to damage nonpermanent 2013). activities in Spain is concentrated. For these reasons it has been a elements (vehicles, cattle, stored goods) or light structures (catwalks or temporary wooden structures), and second, the capacity to destroy completely or partially permanent structural Flood frequency analysis for long scale periods show natural climatic oscillations into so-called Little Ice Age. There are general patterns, affecting most of subject sector secular form large human processes. Recently, tourism elements, either in an urban or in a rural environment: stone bridges, walls and other defensive elements, watermills, buildings, irrigation systems, or roads and railroads. Regarding basins, but also some local anomalies or singularities. To explain these differences and analogies it is not enough to use purely climatic factors. In this way, we and housing construction have introduced new demographic patterns agriculture, a flood is considered destructive if it has rooted out large fields, or if it has destroyed the harvest or the productive plants (grapevines, fruit trees), removing the productive soil analyze human factors that have been able to influence the variability of floods along last 6 centuries (demography, hydraulic infrastructures, urban in the area. and leaving large fluvial deposits of any kind – in summary, catastrophic situations that will need important economic resources and several years for a full recovery, or that mean the development...). This process of land occupation has been on a topographically abandonment of the affected elements. The classification system does not take into account human fatalities due to the occurrence of this kind of impact being random in relation with This approach improves strongly understanding of mechanisms producing major flood events on Eastern coast of Iberian Peninsula, with identification and climate and prone to flood events occurrence of extraordinary range the severity of the flood. evaluation of natural and human factors involved on that. and catastrophic space. C. Historical flood series D. Intra-annual distribution E. Mapping CHRONOLOGIES FROM THE DATABASE E.1 Extraordinaries floods Location Region River Period Years Flood cases Intra-annual distribution Catalunya Sobiráns 1687 -2012 325 70 Sant Adriá Catalunya Besós 1447 -2012 565 25 Besós 13 Balaguer Catalunya Segre/tributaries 1680 -2012 332 21 Barcelona Catalunya Lit.uadis 1389 -2012 623 101 20 Catalunya Capaspre 1755 -2012 257 37 Tarragona Catalunya Francolí 1597 -2012 415 28 El Prat de Catalunya 1517-2012 495 15 Llobregat 94 Cat 20th Girona Catalunya Ter/Onyar 1322 -2012 690 118 % 13th 14th 15th 16th 17th 18th 19th 21th La Seu Catalunya Segre/ 1453 -2012 559 10 Ext d´Urgell tributaries 26 Lleida Catalunya Segre 1156-2012 856 48 Mataró Catalunya 1740 -2012 272 96 Tortosa Catalunya Ebro 1320 -2012 692 46 5 E.2 Catastrophics floods Murcia Murcia Segura 1285 -2012 727 178 Caravaca Murcia Argos 1600 -2012 412 76 Valencia Valencia Turia 1321-2012 691 102 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec TOTAL 15 Cities 1054 flood cases Historical flood data series of Eastern Spanish The intra-annual distribution of flood episodes for the entire period Coast (14th-20th centuries): considered has a typical Mediterranean climate concentration in Cities: 15 late summer months. Extraordinary and catastrophic events are Flood cases: 1054 mainly concentrated in that period, when heavy rain episodes are 13th 15th 16th 18th 19th 20th 21th Catalunya: 698 (66,3%) 14th 17th more frequently recorded. Secondly, there is a minimum during May Valencia: 102 (9,6%) to August associated with those heavy rains episodes those caused Murcia: 254 (24,1%) by rivers with pluvio-nival hydrological regimes. Available period: AD1156-2012

F. Global results G. Local results Population (x1000) Flood cases by decade. VALENCIA Flood cases by decade. LA SEU D'URGELL Flood cases by decade. MATARÓ Flood cases by decade. BALAGUER 5 12 5 120 5 9 16 13 900 The results of this study enhanced the findings of the study Barriendos et. al (2014) conducted locally in the region of Maresme by 1 Cat Cat 800 Cat Cat 14 10 Ext 100 Ext analyzing PREDIFLOOD database. In that study, differences between variability in catastrophic events and the extraordinary are 4 4 Ext 700 Ext Pop 12 Pop noted. The series of catastrophic events show a natural variability of flood events associated with atmospheric dynamics. On the 8 80 600 10 3 3 other hand, the extraordinary variability seems more linked to human processes related to occupation of flood-prone areas. In the 500 6 60 8 study area this occupation has followed an anarchic and chaotic urban planning driven by a growth that accelerated since the mid- 400 2 2 6 20th century due to the massive influx of immigrants and residential speculative processes. 4 40 300 4 200 1 1 2 20 Amplification of PREDIFLOOD with long series of other cities in the Spanish Mediterranean coast (Valencia, Murcia and Caravaca) 2 100 0 0 0 0 0 0 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1800 1891/1900 1991/2000 show that patterns in flood regimes detected in the Maresme are extensible to the rest of the Mediterranean coast because it is 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 noted a similar variability in both, extraordinary and catastrophic events. This validates the method used and makes it extensible Population (x1000) Population (x1000) Flood cases by decade. BESÒS Flood cases by decade. LLEIDA Population (x1000) Flood cases by decade. MURCIA Flood cases by decade. GIRONA 12 400 5 4,5 120 12 100 10 14 to other Mediterranean areas with similar features. 2 6 Cat 4 Cat 350 Cat 90 Cat Ext 10 100 Ext Flood cases by decade Population (x1000) 10 Ext Pop 80 4 3,5 Total flood cases (1285-2012) 45 4000 Pop Ext Pop 300 70 8 16 8 3 80 40 250 3500 60 3 2,5 14 60 6 200 35 6 50 3000 2 40 2 12 150 30 4 1,5 40 4 2500 30 100 10 1 25 20 1 2 20 2 2000 50 10 0,5 8 20 0 0 0 0 0 0 1500 0 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 6 15 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1000 Population (x1000) 4 10 Flood cases by decade. CALELLA Population (x1000) Flood cases by decade. BARCELONA Population (x1000) Flood cases by decade. TARRAGONA Population (x1000) Flood cases by decade. CARAVACA 6 25 6 3 14 8 7 2000 3,5 11 120 15 500 Cat 2 5 Cat 1800 Cat Cat 7 3 Ext 12 5 5 100 Ext Ext Ext 20 1600 0 0 Pop 0 6 Pop Pop Pop 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 2,5 10 1400 4 Ext Cat Pop 4 80

1285 1305 1325 1345 1365 1385 1405 1425 1445 1465 1485 1505 1525 1545 1565 1585 1605 1625 1645 1665 1685 1705 1725 1745 1765 1785 1805 1825 1845 1865 1885 1905 1925 1945 1965 1985 2005 5 1200 15 8 2

3 4 1000 60 3

In the whole study area the number of floods has increased 6 1,5 Floods cases in >=2 cities 800 10 3 2 40 2 12 during the period studied, both as regards extraordinary as 600 4 1 2 catastrophic. Catastrophic floods show three climatic 400 5 1 20 1 10 2 1 0,5 patterns of high frequency not related with urban growing: 200

0 0 0 0 0 0 0 0 8 1580-1620, 1760-1800, 1840-1880.1890-1910 and 1960- 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1 800 1891/1900 1991/2000

1980 May be related to growing urban Population (x1000) Flood cases by decade. ARENYS Population (x1000) Flood cases by decade. EL PRAT DE LLOBREGAT Flood cases by decade. TORTOSA Population (x1000) Total flood cases by city 14 70 4,5 50 6 6 14 8 12 - Extraordinary floods show clear relation with demography, 4 200 Cat Cat 4 Cat 45 12 60 180 including late 20th C. reduction, explanaible by Ext 12 Ext Ext 4 5 40 160 3,5 Pop Pop Pop 140 demographic reduction and efficient hydraulic protection 10 10 50 35 120 3 2 4 100 works. 30 8 8 40 2,5 80 3 25 60 Finally, an upward trend as regards the (catastrophic and 2 0 6 6 30 40 20 1617 1642 1667 1692 1717 1742 1767 1790 1815 1840 1865 1890 191 5 1940 1964 1988 20 extraordinary) floods affecting two or more municipalities 2 1,5 15 0 * 1617, first case that meet the criteria 4 4 20 1 in the same month is noticed. The latter finding could come 10 1 Besós Lleida Urgell El Prat El

2 Calella Girona ´

2 10 Arenys Murcia Mataró Tortosa

0,5 Valencia Balaguer explained by greater exposure or even an accentuation of 5 Caravaca Barcelona Tarragona 0 0 events mesoscale character. 0 0 0 0 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700 1791/1800 1891/1900 1991/2000 1285/1300 1391/1400 1491/1500 1591/1600 1691/1700800 1791/1 1891/1900 1991/2000 1 11 21 31 41 51 61 71 SeuLad

References: Alberola, A. 2006. "Entre la sequía y la inundación. Una aproximación a las avenidas históricas de los ríos valencianos durante el siglo XVIII", in Gérard Chastagnaret and Antonio Gil Olcina (eds.), Riesgo de inundaciones en el Mediterráneo occidental, Casa de Velázquez-Universidad de Alicante, p. 1-30. ISBN 84-95555-89-1. Barriendos, M., Ruiz-Bellet, J.L., Tuset, J., Mazón, J., Balasch, C., Pino, D., Ayala, J.L.: 2014, "The ’Prediflood’ database of historical floods in Catalonia (NE Iberian Peninsula) AD 1035-2013, and its potential applications in flood analysis", Hydrology and Earth System Sciences, 18: 1-17. DOI: 10.5194/hess-18-1-2014. Gil-Guirado, S.: 2013. Reconstrucción climática histórica y análisis evolutivo de la vulnerabilidad y adaptación a las sequías e inundaciones en la Cuenca del Segura (España) y en la Cuenca del Río Mendoza (Argentina). Doctoral dissertation. Universidad de Murcia, Murcia, España. .