The management of hydrological extremes and impacts on society is
inadequately understood because of the combination of short-term
hydrological records, an equally short-term assessment of societal
responses and the complex multi-directional relationships between the two
over longer timescales. Rainfall seasonality and inter-annual variability on
the Pacific coast of Central America is high due to the passage of the Inter
Tropical Convergence Zone (ITCZ) and the El Niño–Southern Oscillation
(ENSO). Here we reconstruct hydrological variability and demonstrate the
potential for assessing societal impacts by drawing on documentary sources from
the cities of Santiago de Guatemala (now Antigua Guatemala) and Guatemala de
la Asunción (now Guatemala City) over the period from 1640 to 1945. City
and municipal council meetings provide a rich source of information dating
back to the beginning of Spanish colonisation in the 16th century. We use almost
continuous sources from 1640 AD onwards, including > 190 volumes
of
Inter-annual precipitation in Central America is highly variable and has extreme socio-economic consequences (Magrin et al., 2014). The region is considered one of the most sensitive tropical regions to climate change, particularly through hydrological cycle impacts (Magrin et al., 2014; Giorgi, 2006). Rainfall anomalies are expected to impact streamflow, freshwater supplies, agriculture and food production (Hidalgo et al., 2013; Magrin et al., 2014). Additionally, extreme weather events lead to natural hazards, such as floods, landslides and droughts, with a cascade of social consequences (Magrin et al., 2014). For example, the strong drought of 2014–2016 has left more than 2 million people in Central America struggling to feed themselves due to poor harvests (WFP, 2015, 2016). At the other extreme, the flooding and winds caused by Hurricane Mitch in 1998 killed more than 11 000 people in this region and caused damage estimated at more than USD 5 billion (Lott et al., 1999). Projections of increased drought in Central America in the next few decades (Magrin et al., 2014) are, however, less certain as a consequence of the short record of 20th century instrumental observations, high rainfall variability and the lack of any significant trend in total precipitation over the last 50 years across the region (Aguilar et al., 2005; Marengo et al., 2014). It is therefore necessary to extend the climate record in order to better understand the mechanisms that govern the hydrological cycle and extreme weather events.
From the 16th to the early 19th century, the Kingdom of Guatemala formed part
of Spain's colonial empire and was governed at the regional and local level by a
hierarchy of Spanish officials, of whom the most senior was the President of
the
The significance of colonial records relating to the activities of city
councils (
Both capitals of the Kingdom of Guatemala, Antigua Guatemala and Guatemala
City, are in the Dry Corridor (DC) of Central America, an ecological area
defined by the Central American dry tropical forest that starts in Chiapas
(southern Mexico), and extends across the central and lowland areas of the
Pacific coast of Guatemala, El Salvador, Honduras and Nicaragua to the Pacific
province of Guanacaste in Costa Rica (Fig. 1; Arias et al., 2012). This
region, which holds most of the Central American population (Imbach et al.,
2015), not only shares a common vegetation, but also a similar climate and
rainfall variability (Fig. 1), presenting lower precipitation than many
other areas of Central America and more severe and prolonged dry periods
(Arias et al., 2012). Thus, the common precipitation regime along the
Pacific coast of Central America (Fig. 1) makes the climatic proxies held in
the
Here we construct a new semi-quantitative hydrological index valid for the
Pacific coast Dry Corridor using the
Precipitation on the Pacific coast of Central America is characterised by a
well-defined rainy season following the annual migration of the Inter
Tropical Convergence Zone (ITCZ, Portig, 1965; Hastenrath, 1967). In
Guatemala, this rainy season is from May to October (Figs. 2 and 3), ending
when the ITCZ migrates south (Portig, 1965; Hastenrath, 1967). During the
Northern Hemisphere summer, easterly waves and tropical cyclones travel
across Central America with the north-east trade winds, bringing
precipitation to the region (Hastenrath, 1967). Storms, or
Hastenrath (1967, 2002) proposed that one of the main mechanisms causing the
MSD is the northward movement in the early summer and the southward movement
in the late summer of the Pacific ITCZ. However, the fact that the time
frame of the MSD remains constant at different latitudes and that the MSD
occurs as far as 20
Average monthly precipitation for Guatemala City Observatory
(INSIVUMEH) between 1960 and 1990 (black, panel
Averaged monthly rainfall and temperature for the cities of Guatemala and the closest meteorological station to Antigua Guatemala (called Suiza Contenta) for the period 1980–2010 (INSIVUMEH). The grey line shows an estimation of the monthly average run-off per month of the Pensativo River for the 1981–2010 period based on results from the ORCHIDEE land surface model. The asterisks show the highest estimated river flow per month and the year in which it happened. All the highest estimated river flow values of the rainy season correspond to La Niña years.
On an inter-annual timescale, the tropical North Atlantic SSTs (Aguilar et al., 2005) and ENSO (Fig. 2; Magaña et al., 2003) play significant roles in the precipitation of the Pacific coast of Central America. Positive North Atlantic SSTs enhance local evaporation and consequently increase water vapour transport to Central America by the easterly trade winds (Aguilar et al., 2005; Hastenrath and Polzin, 2013; Polzin et al., 2015). SSTs in the tropical Atlantic can also increase precipitation over Central America by strengthening the Atlantic TC season (Aguilar et al., 2005). On the other hand, positive SSTs in the eastern tropical Pacific associated with El Niño decrease rainfall over the Pacific coast of Central America (Fig. 2; Magaña et al., 2003; Hastenrath and Polzin, 2013; Polzin et al., 2015). Warm ENSO conditions lower the number of tropical cyclones over the North Atlantic basin due to an increase in vertical wind shear and weaker easterly waves (Gray, 1984; Goldenberg and Shapiro, 1996; Goldenberg et al., 2001). El Niño also strengthens the easterly Caribbean low-level jet due to a negative SST gradient from the Pacific to the Atlantic, which results in an enhanced pressure gradient over Central America (Enfield and Alfaro, 1999; Rauscher et al., 2008, 2011) and more subsidence along the Pacific coast (Magaña et al., 2003). In addition, El Niño years are related to a southward position of the ITCZ (Waliser and Gautier, 1993) due to a weaker SST meridional gradient between the eastern Pacific cold tongue region and the Mexican coast warm pool, bringing drier conditions to this region (Magaña et al., 2003). The strongest reduction in precipitation during El Niño occurs during the MSD prior to the peak of a warm ENSO event, which usually occurs during boreal late autumn and winter (NDJ; Fig. 2; Trenberth, 1997; Curtis, 2002). Conversely, La Niña increases rainfall over the Pacific coast of Central America due to the opposite factors mentioned above (Fig. 2; Curtis, 2002; Magaña et al., 2003).
The cities of Santiago de Guatemala (Antigua Guatemala) and Nueva Guatemala de la Asunción (Guatemala City) are located 35 km apart in the south-central region of the Republic of Guatemala at an altitude of approximately 1500 m (Fig. 1). The cities share a similar temperate climate and have an annual average precipitation of approximately 1100 mm for Antigua Guatemala and 1200 mm for Guatemala City, with monthly distribution following the bimodal pattern of the Pacific coast of Central America (Fig. 3). The Pensativo basin (Fig. 1), circumscribed on the east side by the Guacalate basin and part of the bigger Achiguate river catchment where Antigua Guatemala is located, is characterised by steep slopes and sediments with high permeability (Quiroa-Rojas, 2004). Consequently, the river flow is intermittent and typically ceases during the dry season (Fig. 3). High discharge rates and sediment load in response to the rainy season have caused regular problems for the population in the area of Antigua Guatemala, including severe flooding (Quiroa-Rojas, 2004).
These two cities, which were successive capitals of the Kingdom of Guatemala, contained a significant proportion of the Spanish population in Central America during the period of study (MacLeod, 2010). Santiago de Guatemala (Antigua Guatemala; MacLeod, 2010) was founded in 1543 after a strong mudflow from the Agua volcano in September 1541 destroyed the previous capital, today known as Ciudad Vieja (Hutchison et al., 2014). Santiago became the most important city in Central America (Lutz, 1997) and after 1773, following an earthquake of an inferred magnitude of 7.5 (Villagran et al., 1996) that destroyed most of the city, the capital was relocated again, this time to the Valley of the Ermita, and named Nueva Guatemala de la Asunción (today Guatemala City).
Minutes of the meetings of the city and municipal councils (
Direct information in the
A key additional source of climatic information during the colonial period
is the rogation ceremonies carried out by municipal and church authorities.
A rogation ceremony, known as a
Rogation records as indicators of hydrological stress in Guatemala need
cautious analysis. After the first quarter of the 18th century, as
noted in the
In assessing the information contained in the
The Claxton (1986, 1998) compilation of drought and flood episodes in
Guatemala starting from the middle of the 16th century is used in this
study to corroborate the information obtained from city and municipal
council records and to fill gaps associated with the way in which the
information was recorded in the
The climatic information contained in the
“… we approach the time [of year] when this Illustrious City
Council customarily celebrates a
If the rainfall season was wetter than usual, the
“… the measures that may be taken to repair the bridge of Los Esclavos, which appears to have been ruined as a result of the winter [weather] and the copious amounts of water that have flowed through [the bridge] …” (AGCA, 14/11/1727, sig. A1.2.2, leg. 1791, Exp. 11785, folio 49).
In the most extreme cases, such as severe flooding, the
“This council [meeting] discussed the news … from the towns of San Juan Gascón and San Miguel el Alto … to the effect that the winter season having been so harsh this year, and the floods severe, a great portion of the main [water] tank collapsed at the point where two streams converge and flow [into it]; it was agreed that the tank should be repaired …” (AGCA, 08/10/1697, sig. A1.2.2, leg. 1785, Exp. 11779, folio 117).
However, significant flooding events also occurred during other periods of the rainy season, such as in 1878:
“[Heavy] rain has caused serious damage along the Pensativo River. The
riverbed having silted up it flowed over properties situated along its
banks. Establishing the means to prevent such damaging events [in the
future] is a matter of the utmost importance; accordingly … an
open meeting of the municipality will be called to identify the most
appropriate measures to address this ongoing threat …” (AHMAG
10/07/1878,
Prior to the 1720s, if rainfall during the first part of the rainy season was irregular and lower than usual, the city celebrated a rogation ceremony to the Virgin of Our Lady of Socorro in May. On occasion these records also include information about the state of fields and crops:
“[The meeting] discussed the fact that although the season for rain is well advanced, … we have experienced such a lack of it that as of today it has not rained [at all], resulting in heat so excessive that harmful consequences are feared, and that lack of water will also lead to poor harvests, assuming there are any. It was [in consequence] resolved … that a [service of] rogation would be held before the most holy image of Our Lady of Socorro, in her chapel in the Holy Cathedral Church of this city on Saturday 25th of this month”(AGCA, 23/05/1709, sig. A1.2.2, leg. 1787, Exp. 11781, folio 100).
If the situation was particularly severe, a rogation ceremony to Our Lady of Socorro was carried out in the middle of the rainy season (July, August or September), as the following example from 1699 shows:
“In this council [meeting] it was agreed that due to the lack of rain that … is causing the severe [unspecified] illnesses which are afflicting this city, … and which also threatens the failure of the critical wheat and maize harvests … the city magistrates will request … licence to hold a procession of rogation to the most holy image of Our Lady of Socorro, [the procession to end] in the church of the convent of our Lady of Mercy. To this end, they are to … make … arrangements … for the procession … as well as for a sung Mass to be celebrated on the day following …”(AGCA, 28/07/1699, sig. A1.2.2, leg. 1785, Exp. 11779, folio 206).
As rogation ceremonies incurred substantial costs, they were only conducted
when drought was severe and the impact significant (Martin-Vide and
Vallvé, 1995). These ceremonies are much more common in the record
before the tradition of the
Classification of hydrological conditions for the cities of Guatemala and Antigua Guatemala.
On occasion, information about crop shortages, harvest failures or
flooding events is recorded in the first few months (dry season) of the
following year's
A five-point scale (from
Accounts of droughts and inadequate rain from Claxton (1986, 1998) and
Pardo (1944) were also used to reconstruct five dry years (1694, 1747, 1752,
1803, 1810; between 1640 and 1840) in which crop shortages are mentioned in the
Following Kelso and Vogel (2007), we have given a confidence rating to each
year in our reconstruction. The confidence rating ranges from 1 to 3 based
on both the number of entries recording an extreme weather event in the
The instrumental records from three meteorological stations located in the vicinity of Antigua Guatemala and Guatemala City and held by the Instituto Nacional de Sismología, Vulcanología, Meteorología e Hidrología (the National Institute of Seismology, Volcanology, Meteorology and Hydrology) are compared with the overlapping period in the rainfall reconstruction (1910–1945). Two sets of rainfall records are from the area of Antigua (El Potrero, 1910–1971; Antigua, 1934–1971) and the third is located in Guatemala City (INSIVUMEH, 1928–1912). An additional six years from 1879–1882 and 1896–1897 reported in Guatemala City are also compared with respect to the reconstructed hydrological index. Estimated monthly run-off data per month for the period 1981–2010, derived from the land surface model ORCHIDEE and based on precipitation observations, demonstrate the highly seasonal flow of the Pensativo River (Fig. 3).
Five-category semi-quantitative hydrological index for the cities
of Antigua Guatemala and Guatemala City starting in 1640 and ending in 1945.
Rainfall in each year is classified as very dry
The books of the
(
Dry conditions persisted in the 1640s–1740s, 1820s and 1840s (Fig. 5),
including some exceptionally dry decades when 4–5 years were recorded as dry
(1660s, 1720s, 1730s and 1820s; Fig. 5). These were linked in the
“This [meeting] … discussed the fact that due to the lack of rain [the city] is experiencing an intense heat harmful to public health, such that a widespread epidemic had to be feared, as lack of pasture was leading to the death of livestock, and it was also reported that the crops had not borne fruit. [Thus] … it was resolved … that the city would appeal to the Holy Virgin Mother of God … carrying out a rogation ceremony to and procession of the miraculous image of Our Lady of Socorro” (AGCA, 19/05/1701, sig. A.1.2.2, leg. 1786, Exp. 11780, folio 22–23).
Conversely, the 1760s–1810s seemed to be dominated by wetter conditions characterised by a low number of dry years and a relatively high number of wet years, including some major flooding events or crop shortages due to an abundance of rain, such as in 1762, 1783, 1792 and 1816:
“The copious rains have been the reason why, in various districts, maize harvests have been scarce … and why it is feared, with good reason, that this capital will suffer shortages of said grain. This calamity should have and could have been anticipated, had regional officials been vigilant in monitoring the situation and made provision for bringing more land under cultivation which is their primary obligation; but they have not complied, nor have they even informed this government of the state of the harvest in their respective regions …” (AGCA, 02/1817, sig. A1.2.2, leg. 2192, Exp. 15743, folio 55–56).
The 1780s and 1810s were the decades with the highest number of wet years (three and four wet years per decade respectively), whereas the 1690s and 1890s are notable for their high variability between extremes of both drought and flooding (Fig. 5).
In the overlapping 35-year period (1910–1945; Fig. 6) the driest (1914) and
wettest (1933) years of the 20th century recorded by three nearby
meteorological stations are correctly classified in the documentary-based
hydrological reconstruction (Fig. 6). A further two dry years (1912 and
1920) and two wet years (1921 and 1944) also match the anomaly recorded from
instrumental observations, but two years (1911 and 1923) categorised as
wetter than average in the index were drier than average according to the El
Potrero rainfall observations. In addition, the wetter conditions of
1896–1897 according to instrumental observations from Guatemala City are
picked up in the
A near-continuous reconstruction of hydrological variability from 1640
onwards for the Pacific coast of Central America was derived from
documentary sources collected from the Archivo General de Centro América (AGCA, Guatemala City) and the
Archivo Histórico de la Municipalidad de Antigua Guatemala (AHMAG Antigua Guatemala). In addition the
Number of dry and wet events per decade based on information from city and municipal council records; Claxton (1986, 1998) and Pardo (1944).
During the 17th and 18th centuries when dry years were much more
common than in later centuries (Fig. 5), the city council arranged rogation
ceremonies to pray for good rain (see Sect. 4.1; AGCA, 19/05/1701, sig.
A.1.2.2, leg. 1786, Exp. 11780, folio 22–23). In addition, as further
mitigation measures, the city council ordered second sowings to be made in
other regions with a more favourable climate, such as Escuintla nearer to
the Pacific coast, where both rainfall (average precipitation for the city
of Escuintla is 2800 mm) and temperatures (annual average temperature of
26
“The
Semi-quantitative hydrological index reconstructed in this study (1910–1945) and annual precipitation anomalies (calculated with respect to the average precipitation for the 1934–1945 period) for the meteorological stations of Antigua Guatemala, Guatemala City and El Potrero.
In contrast, wet weather events were commonly associated with flooding in
the city of Antigua Guatemala by the Pensativo River. It was recognised that
Pensativo floods were not only caused by intense rainfall, but also
the anthropogenic activities in the river catchment, specifically the clearing
and planting of crops on surrounding hillslopes, leading to soil erosion and
the
silting up of the riverbed. In response, the city council of Antigua
ordered preventative annual dredging of the Pensativo River prior to the
rainy season, as illustrated by an entry from 1877 (frequently repeated)
calling on the authorities to arrange for “this important work” to be
undertaken and to make owners responsible for dredging stretches of the
river bordering their lands (AHMAG 13/04/1877,
In addition, the municipal authorities of Antigua Guatemala prohibited the
clearing and planting of hillslopes near the Pensativo River. However,
the
However, in years when these prohibitions were enforced in Antigua and the surrounding towns, damage to the city associated with intense wet events was reduced or even avoided, as this example from 1923 shows:
“… the Council agreed that this municipality should inform the
Ministry of Government and Justice that notwithstanding the exceptionally
strong storm that struck on 11, 12 and 13 October, which wreaked havoc
across the country and beyond, this city suffered no damage from the
Pensativo River, the bed of which did not become silted up as it has in
times past. This confirms, and as shown by experience, that the prohibition
on clearing and planting on the slopes of the hills that surround this city
… has been highly beneficial” (AHMAG 13/11/1923,
The accounts presented above demonstrate that the society of the Pacific
coast of Central America was conscious of the societal impacts of extreme
hydrological events and had identified successful mechanisms to remediate
the risk. In addition, the
Instrumental observations should be used to calibrate any proxy climate
reconstruction, including document-based hydrological records (e.g. Rodrigo
et al., 1999; Rodrigo, 2008), and an overlap of 30 years for calibration and
30 years for verification has been recommended (Rodrigo et al., 1999).
However, we were unable to find any instrumental river flow data for the
Pensativo during the reconstructed time period, and the earliest continuous
instrumental precipitation record near Antigua Guatemala starts in 1910
(El Potrero). The document-based index is relatively consistent with these
meteorological station precipitation records (1910–1945, i.e. 35-year
overlap; Fig. 6) despite the fact that the information inferred from the
A discrepancy arises in the years 1911 and 1923, which were categorised as wet in the
index but according to the El Potrero rainfall observations were drier
than average. In 1923, the El Potrero monthly rainfall data show that
precipitation between May and August was less than half the average for the
1910–1970 period, but rainfall accumulation in October was 3 times
(325 mm) the average (119 mm) due to a
In the second case, precipitation for all months of the 1911 wet season recorded at El Potrero was below average, with the exception of May 1911 (198 mm with respect to a monthly average of 99 mm). However, the municipal council minutes highlight recorded problems in the Pensativo River during September:
“The Alcalde presiding informed the Corporation that, as a consequence of the
heavy rains of yesterday … the Pensativo River overflowed in the
region of the bridge of San Ignacio, and that he thinks that the main damage
has been to a rocky section of the riverbed in the
While heavy rains over the late part of the rainy season in association with saturated soil from previous rainfall during the wet season would explain a higher flow in September (see monthly average flow in Fig. 3) and account for the above report in the minutes of the meetings of the municipal council, the monthly precipitation from El Potrero, in contrast, indicates well below average rainfall in July 1911 (46 mm compared to the monthly average of 150 mm), August 1911 (48 mm compared to the monthly average of 130 mm) and September 1911 (91 mm compared to the monthly average of 208 mm). Either very localised and concentrated convective rainfall occurred within the Pensativo catchment and did not affect El Potrero, or there was a problem with the instrumental records.
Of course, it is also possible that some of the dry years according to the
instrumental record reflect a deficit in precipitation (a
meteorological drought) that was minor enough not to cause the serious
societal consequences characteristic of an agricultural drought induced by
soil water deficiency and plant water stress. Climatic information was
recorded in the
Since the Pensativo River has been a persistent threat to the city of
Antigua Guatemala, measures of overabundance of rainfall are less
susceptible to changes in the way information is reported in the documentary
records, especially for years classified as
“The … Alcalde presiding informed the Council that because the
riverbed of the Pensativo had silted up, the bridge between Pavón and
San Ignacio is now useless, as the water flows over the bridge …
as its repair is vital, he requests authority to proceed [noting that] some
citizens have offered to contribute [towards costs]. The …
Corporation agreed that a budget for the works should be submitted to the
next meeting …” (AHMAG 20/09/1932,
Although changes in the way the climatic information was recorded in the
The persistent dry period of the first part of the reconstruction, from the 1640s to the 1740s, contains the highest number of dry years of the reconstruction, peaking in the 1720s–1730s (Figs. 4 and 5). The number of wet years during this period was also lower, with the exception of the 1690s when the number of both dry and wet years was extremely high compared with the rest of the period of study. Additionally, of the 32 years between 1640 and the 1820s when we know rogation ceremonies were celebrated, 29 occurred during this early period; 16 of the rogation ceremonies took place between the end of April–May, implying an irregular start of the rainy season, while the other 13 were made during the onset of the rainy season (July–September), implying lower precipitation during the early part of the rainy season and stronger drought conditions over the MSD. These specific conditions are characteristic of El-Niño-related rainfall anomalies in the Pacific coast of Central America (Fuentes et al., 2002; Curtis, 2002) and lead to lower annual averaged precipitation (Magaña et al., 2003).
We tested this relationship by comparing our reconstruction with the proxy-based ENSO reconstruction of Gergis and Fowler (2009) and found that 35 % of dry years coincided with a reconstructed El Niño year compared to 26 % with La Niña. The years in which both the early and late rogation ceremonies were undertaken were no more likely to match a reconstructed El Niño year than a La Niña, and more early rogation ceremonies were celebrated in the years identified by Gergis and Fowler (2009) as La Niña. These results are contrary to expectations based on recent observations that warm ENSO conditions are coincident with lower precipitation, an irregular start of the rainy season and a stronger midsummer drought in the Pacific coast of Central America (e.g. Fuentes et al., 2002; Curtis, 2002; Magaña et al., 2003). The non-stationary relationship of regional rainfall teleconnections to ENSO over the multi-decadal to centennial timescale is now well recognised (e.g. Stahle et al., 1998; Hendy et al., 2003; Adamson and Nash, 2014; Ashcroft et al., 2016). In addition, the Gergis and Fowler (2009) ENSO time series is also based on a restricted number of proxy records and heavily dependent on the persistence of teleconnections. It is important that the analysis of historical sources be expanded to improve the available reconstructions of past ENSO events.
The persistent dry period from the 1640s to the 1740s was coincident with
the period of cooler Northern Hemisphere (NH) temperatures called the Little
Ice Age (LIA), which started in the 1500s and ended in the 1800s (e.g.
Bradley and Jones, 1993; Jones et al., 1998; Mann et al., 1998, 1999), and
significant glacier advances, including a maximum extent in the South
American Andes between the 1630s and the 1730s (Jomelli et al., 2009). This dry
signal has been recognised at other sites in Mesoamerica, including from
documentary sources in Mexico (e.g. Florescano, 1980a, b; Florescano et
al., 1980; Swan, 1981; Endfield and O'Hara, 1997). Several authors have
suggested an increase in frequency of warm ENSO conditions during the LIA
(Cobb et al., 2003; Langton et al., 2008). In addition, strong cooling over
the NH, including the North Atlantic basin, may have been responsible for the
southward movement of the ITCZ during the LIA by an estimated
The oscillations of the ITCZ (Portig, 1965; Hastenrath, 1967, 2002) and the Pacific (e.g. Fuentes et al., 2002; Magaña et al., 2003) and North Atlantic SSTs (e.g. Aguilar et al., 2005) are the dominant mechanisms that control precipitation over the Pacific coast of Central America. Therefore, lower equatorial North Atlantic SSTs (Haug et al., 2001), including the Caribbean region (Winter et al., 2000) and a southward movement of the ITCZ caused by a strong cooling over the NH during the LIA (Haug et al., 2001; Sachs et al., 2009), may have been responsible for the dry conditions that prevailed during the first 100 years of this reconstruction across the Pacific coast of Central America. This hypothesis is further supported by the increase in tropical Pacific sea surface salinity (Hendy et al., 2002) and the strengthening of the South American summer monsoon (SASM) during the LIA associated with a southward position of the ITCZ found in observation and modelling studies (e.g. Bird et al., 2011; Vuille et al., 2012; Apaéstegui et al., 2014). In addition, the possibility of dominant El Niño conditions during the LIA (Cobb et al., 2003; Langton et al., 2008) may also have played a role, although the question of how ENSO varied through the LIA is still debated (Yan et al., 2011; Henke et al., 2017; Emile-Geay et al., 2013).
From the 1750s, only 3 decades are marked by a larger number of years
with dry conditions, specifically the 1820s, 1840s and the 1890s, and there
is a period (1760–1810s) during which wet conditions dominate in the
reconstruction. This shift to fewer droughts recorded in the
The changes in the frequency, extent and severity of extreme weather events (floods and droughts) found in our hydrological index provides valuable information for improving our understanding of Central America's inter-annual and inter-decadal rainfall variability and its sensitivity to climate change. The most plausible explanation for the shift to wetter conditions and a lower number of dry years from the 17th century to the 19th century is the end of the Little Ice Age, a possible tendency toward more La Niña events, warmer equatorial North Atlantic SSTs and a northward movement of the ITCZ. On the other hand, global warming projections suggest that the ITCZ will move southward again by the end of the century, probably related to warm ENSO-like conditions within the eastern tropical Pacific (Rauscher et al., 2008; Hidalgo et al., 2013). This scenario could create conditions similar to the 1640s–1740s because a stronger Caribbean low-level jet and low-level easterlies would cause surface divergence in conjunction with an earlier westward amplification and intensification of the North Atlantic subtropical high, which would bring drier conditions to the Pacific coast of Central America (Rauscher et al., 2008). An increase in the frequency of widespread droughts and an earlier start and intensification of the MSD of up to 25 % is forecast in some areas of Guatemala, Honduras and El Salvador (Neelin et al., 2006; Rauscher et al., 2008, 2011; Hidalgo et al., 2013). If drier conditions in the future mimic those experienced during the 17th and 18th centuries, as recorded in city and municipal council records, it would not be surprising to see an increase in agricultural problems, food shortages and malnutrition for the most vulnerable part of the population. In this climatic scenario, the Central American population would not only once again suffer the consequences of widespread droughts, but this time they could be accentuated by the anthropogenic changes made to the land surface over this region affecting the evapotranspiration, infiltration, surface run-off, water storage and development of drought conditions (Van Loon et al., 2016). Therefore, more than 300 years after the 1640s–1740s dry period, the scenario projected by general climate models as a consequence of global warming could have stronger socio-economic consequences for the population of this region than in the past, increasing the previous and current Central American high vulnerability to climate change.
The
The supplement related to this article is available online at:
The authors declare that they have no conflict of interest.
This article is part of the special issue “Climate of the past 2000 years: regional and trans-regional syntheses”. It is not associated with a conference.
This is a contribution to the PAGES 2k Network. Past Global Changes (PAGES)
is supported by the US and Swiss National Science Foundations. This work was
supported by a joint University of Bristol postgraduate scholarship and La
Caixa Scholarship to Alvaro Guevara-Murua and by a RCUK Academic Fellowship to Erica J. Hendy. This
work was implemented as part of the CGIAR Research Program on Climate
Change, Agriculture and Food Security (CCAFS), which is carried out with
support from CGIAR Fund donors and through bilateral funding agreements. For
details please visit