Impact of the 2010–2011 La Niña phenomenon in Colombia, South America: The human toll of an extreme weather event

Abstract

The 2010–2011 La Niña (positive phase of El Niño) phenomenon affected four million Colombians, ∼9% of the total population, and caused economic losses of approximately US $7.8 billion, related to destruction of infrastructure, flooding of agricultural lands and payment of government subsidies. We analyzed the spatial patterns of effects on the population, measured as the number of affected persons in each municipality normalized to the total municipal population for 2011, using global (Moran's I index) and local (LISA) spatial autocorrelation indicators, and multiple regression analyses (OLS and ML spatial error model). The spatial autocorrelation analysis revealed two regional clusters or “hotspots” with high autocorrelation values, in the lower Magdalena River Valley (Caribbean plains) and lower Atrato Valley (Pacific lowlands). The regression analyses emphasized the importance of the spatial component as well as the variables related to hazard exposure and social vulnerability. Municipalities in “hotspots” show: (1) a high degree of flooding, as they are located on the Magdalena and Atrato River floodplains, and (2) high social vulnerability, suggested by low values of the ICV (national living conditions index).

Introduction

Climate patterns have changed throughout Earth's history. Since the late 1800s, these changes have been largely caused by increasing amounts of anthropogenic greenhouse gases in the atmosphere. The average temperature of the planet has increased 0.74 °C over the last century, and most of this increase has occurred in the last three decades (Arguez, 2007; IPCC, 2007). It is estimated that increases in the concentration of greenhouse gases will cause additional warming of 1.1–6.4 °C by the end of this century (IPCC, 2007). The increase in global average temperatures is expected to cause increases in extreme weather events, which will, in turn, have effects on ecosystems and society. Such events drive greater changes in natural and social systems than do average climate conditions as a consequence of damage to infrastructure and agricultural lands, diminished ecosystem function, and human death, injury and displacement (Parmesan & Martens, 2008; Parmesan, Root, & Willig, 2000). Climate change, particularly extreme weather events, poses risks and challenges for society. Most research, however, has addressed the climate component of climate change, whereas its impact on human well-being remains poorly understood (NRC, 2009). In social terms, effects of extreme events are evaluated by analyzing the vulnerability of exposed communities. Impacts on socioeconomic systems are often amplified by factors such as social inequality, disease and social conflict. Understanding vulnerability and how it relates to climate change, particularly extreme weather events, is an initial step in managing climate change risks. Geographically explicit vulnerability analysis is critical to understand how interactions between the physical environment and humans change over space and time (Emrich & Cutter, 2011; Montz & Tobin, 2011; Moser, 2010).

Colombia experienced a strong El Niño Southern Oscillation (ENSO) cold phase known as La Niña, from 2010 to 2011. The weather event affected approximately four million people as of September 2011 and caused losses of more than US $7.8 billion, as a consequence of destruction of infrastructure, flooding of agricultural lands and payment of government subsidies (Redacción, 2010a, 2011a). A wealth of data was generated by government agencies and non-governmental organizations on the effects of this phenomenon. Furthermore, such information was used to develop mitigation plans. Participating institutions included the National Office for Disaster Risk Management (Unidad Nacional para la Gestión del Riesgo de Desastres – UNGRD), National Department of Statistics (Departamento Nacional de Estadística – DANE), National Institute of Hydrology, Meteorology and Environmental Studies (Instituto de Hidrología, Meteorología y Estudios Ambientales – IDEAM), the National Geographic Institute (Instituto Geográfico Agustín Codazzi – IGAC) and non-governmental entities such as iMMAP and the United Nation's Office for the Coordination of Humanitarian Affairs (OCHA). Although these institutions presented their data in a spatial format (i.e. maps), rigorous geographical analysis was not done, largely because of time constraints. In this study, we assessed the spatial patterns of ENSO effects on the human population in Colombia, and explored the relationship between such patterns and physical geographic and socioeconomic variables. We first summarize the effect of ENSO on Colombian river flow dynamics and follow with a spatial analysis of the 2010–2011 La Niña event. We conclude with a discussion of our findings.

In Colombia, the annual hydrologic cycle is controlled by oscillation of the inter-tropical convergence zone, superimposed on regional patterns caused by orographic influence of the Andes, evapotranspiration in the Amazon Basin, continent-atmosphere interactions and dynamics of the western Colombian wind currents (Western Colombian Jet – Chocó Jet) (Mesa, Poveda, & Carvajal, 1997; Poveda, Jaramillo, Gil, Quiceno, & Mantilla, 2001; Poveda & Mesa, 2004) (Fig. 1). Over longer time scales, major hydrologic anomalies are experienced during both phases of ENSO (Aceituno, 1988; Poveda, 2004; Poveda et al., 2001) and other macro-climatic phenomena such as the North Atlantic Oscillation (NAO) and Pacific Decadal Oscillation (PDO) (Mesa, Poveda, &Carvajal, 1997; Poveda et al., 2002).

The ENSO warm phase (El Niño) causes droughts in the western margin of Central America, Mexico, the Amazon Basin, northern South America (i.e. Colombia and northeastern Brazil), whereas it produces excess precipitation in the eastern region of Central America, and increased summer rainfall in the Paraná Basin and the Andes of Peru, Bolivia and Chile (Capel, 1999). In Colombia, ENSO has a strong effect on precipitation, river discharge and soil moisture (Montealegre & Pabón, 1992; Poveda & Mesa, 1996; Poveda et al., 2001, 2002; Puertas & Carvajal, 2008; Restrepo & Kjerfve, 2004). The warm phase is associated with an increase in the average air temperature, a decrease in soil moisture and evapotranspiration, a decrease in rainfall and a consequent decrease in the average flow of the rivers in the western, central and northern regions of the country (Poveda et al., 2001). The opposite pattern is observed during the cold phase (La Niña), which is mainly characterized by intense and abundant rainfall, increased river flow and subsequent flooding (Poveda & Mesa, 1996; Mesa et al., 1997; Poveda et al., 2001). ENSO events, however, differ in intensity and spatial extent, so their effects on hydro-climatology are event-specific (Poveda, 2004).

A common variable used to assess the strength of a particular ENSO event is the Southern Oscillation Index (SOI). It is calculated as the normalized difference in surface air pressure between Darwin, Australia (Western Pacific) and Tahiti, French Polynesia (Eastern Pacific). A positive index points to low pressures in the western tropical Pacific and indicates the occurrence of the cold phase (La Niña). A negative index signals the presence of the warm phase (El Niño). According to this index, there were at least 19 El Niño and 17 La Niña events between 1950 and 2011 (NOAA, 2011). Because of their intensity and duration, the warm events in 1957–1958 (8 months), 1965–1966 (12 months), 1972–1973 (10 months), 1976–1978 (18 months), 1982–1983 (14 months), 1986–1987 (16 months), 1991–1992 (17 months), 1997–1998 (12 months) and 2009–2010 (11 months) are notable. Strong cold events took place in 1954–1957 (20 months), 1970–1971 (14 months), 1973–1974 (13 months), 1975–1976 (12 months), 1988–1989 (14 months), 1998–2000 (24 months), 2007–2008 (10 months) and 2010–2011 (10 months) (Fig. 2). Climatic, hydrological and oceanographic disturbances related to these events had dramatic global socioeconomic and environmental repercussions (Capel, 1999).

In Colombia, the 1982–1983 ENSO stimulated scientific and academic interest because of its environmental impacts, particularly in the marine sector (Alvarado, Duque, Flórez, & Ramírez, 1986). Interest only became widespread after the 1991–1992 event, which caused a large decrease in precipitation and Andean river streamflows, and led to a collapse of the national hydropower system (Mesa et al., 1997; Montealegre & Pabón, 1992). The relationship between ENSO and river flow in Colombia was studied by Mesa et al. (1997) and Restrepo & Kjerfve (2000). They showed that ENSO has an earlier and stronger effect on rivers in western, northern and central Colombia, in contrast to a later and reduced effect on rivers in the eastern and southeastern regions of the country. For instance, ENSO explains up to 64% of the inter-annual variability in discharge of the Magdalena River, the main river draining the Colombian Andes (Restrepo & Kjerfve, 2000). Abrupt changes in river discharge have occurred during the past 12 years, and all were related with ENSO cold conditions (Fig. 3a). Wavelet analysis, however, reveals that the contribution of ENSO to flow variability has not been constant over time (Fig. 3b). Caribbean river discharge also reflects the effect of ENSO (Fig. 4). Nevertheless, it is difficult to separate the influence of climate variability from that of anthropogenic disturbance (Restrepo & Restrepo, 2005).

The 2010–2011 ENSO cold event was one of the most intense, in both duration and magnitude (Fig. 2). In 2010, there was a rapid transition between the warm and cold phases of ENSO. Completion of the 2009–2010 warm event was marked by negative SOI anomalies during the first quarter of 2010. Beginning in July, the positive anomalies were consolidated, which initiated the cold event and lasted for 18 months, until December 2011. During that period, the anomalies ranged from 1.9 to 5.2. The only comparable anomalies were observed in the cold events of 1970–1971, 1975–1976 and 2007–2008.