The Pacific and Caribbean Rivers of Colombia: Water Discharge, Sediment Transport and Dissolved Loads

Abstract

Although the South American continent includes three of the largest river basins in the world, the Amazon, the Orinoco, and the Paraná, with some of the highest discharges and sediment loads, a number of comparatively smaller systems in Colombia carry a significant share of the sediment load from the continent. Our objective is to synthesise the role and contribution of river systems in Colombia.

Full text

Chapter 14
The Pacific and Caribbean Rivers of Colombia:
Water Discharge, Sediment Transport and
Dissolved Loads
J. D. Restrepo · B. Kjerfve
14.1
Introduction
Although the South American continent includes three of the largest river basins in
the world, the Amazon, the Orinoco, and the Paraná, with some of the highest dis-
charges and sediment loads, a number of comparatively smaller systems in Colombia
carry a significant share of the sediment load from the continent. Our objective is to
synthesise the role and contribution of river systems in Colombia.
South America measures 17.8 ×106 k m2 and accounts for 12% of the global land sur-
face. However, the continent delivers a disproportionally larger water discharge and
suspended sediment load into the oceans as compared to its area. The three largest
rivers only deliver 7300 km3yr–1 or 24% of the global water runoff. With respect to
suspended sediment load, the South American continent contributes 13% of the glo-
bal load into the oceans. Although most of the discharge and sediment load are due to
the Amazon, Orinoco, and Paraná Rivers (Milliman 1990; Milliman and Syvitski 1992),
the Magdalena River, which empties into the Caribbean Sea (Fig. 14.1), transports more
sediment than either the Orinoco and Paraná Rivers (Milliman and Meade 1983), al-
though it has much smaller water discharge and drainage area.
In general, the drainage basins on the eastern side of South America are large,
whereas the numerous basins with discharge into the Pacific are comparatively small
because of the crowding of the drainage basins west of the Andes imposed by regional
geology and tectonics (Kellog and Mohriak 2001) (Fig. 14.1). However, the discharge
of suspended sediment into the oceans from many smaller Colombian rivers may have
greater impact on the world sediment budget than previously thought (Milliman 1990;
Milliman and Syvitski 1992). Further, land use in South American basins is changing
rapidly and is seemingly causing changes in water, sediment, and nutrient transports,
resulting in regional impacts. From the perspective of ecology and to understand
change impact, there is a need to understand better the geochemistry of these rivers
(Richey et al. 1991).
Water discharge, sediment load, and physical characteristics for the major Pacific
and Caribbean rivers of Colombia have been reported during the past few years
(e.g. Restrepo and Kjerfve 2000a,b, 2002). However, until now data for the dissolved
load of these rivers have not been reported.
Due to the importance of Colombian rivers in the global budgets (Milliman and
Meade 1983; Milliman and Syvitski 1992; Restrepo and Kjerfve 2000a,b), we synthesise
data on water discharge, sediment load, and dissolved load of the principal rivers along
the Pacific and Caribbean coasts of Colombia, make comparisons to other major flu-
vial systems draining into the Atlantic Ocean and elsewhere, and present some result-

J. D. Restrepo · B. Kjerfve170
ing environmental implications and impacts along the Caribbean. This information
is a significant addition to the understanding of (1) chemical weathering processes
occurring on a regional scale, (2) the nature of the organic matter from autochtho-
nous and allochthonous sources, (3) fluvial fluxes to oceans and the context of Colom-
bian rivers in the global budgets, and (4) human impacts on continental and coastal
water systems.
14.2
Water Discharge and Sediment Load
14.2.1
The Pacific Rivers
Climate, geology, relief, and size of the drainage basin are critical factors that deter-
mine river discharge. The basins of the Pacific coast of Colombia, measuring 76365 km2
and extending from latitude 00°36' N to latitude 07°45' N and from longitude 75°51' W
to longitude 79°02' W, are characterised by the presence of active fault systems, high
precipitation rates, slopes frequently steeper than 35°, and dense tropical rain forests.
These conditions are favourable for the occurrence of rapid mass wasting, caused by
slope erosion processes and thus high sediment loads. The basins extend inland 60–
150 km and comprise all of Colombia west of the Cordillera Occidental of the Andes,
linking Panama and Ecuador. They consist of a broad coastal plain and the western
Fig. 14.1. Map showing the major basins in South America draining into the Atlantic Ocean and the
smaller Pacific and Caribbean basins of Colombia

171CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
slopes of the Cordilleras. The principal rivers from north to south are the Baudó, San
Juan, Patía, and Mira (Fig. 14.2).
The Pacific basins are located within the humid tropics, characterised by high but
relatively constant temperature, high rainfall rates, and high humidity. Average rain-
fall ranges from 2000 to 12 700 mm yr–1 (Eslava 1992). The rainfall distribution is bi-
Fig. 14.2. Map of the Pacific and Caribbean coasts of Colombia showing the principal rivers, the main
drainage basins indicated in Tables 14.1 and 14.2, the Western, Central, and Eastern Cordilleras (solid
triangles), and the estimates of total sediment load and, accordingly, sediment yield into the Pacific Ocean
and Caribbean Sea

J. D. Restrepo · B. Kjerfve172
modal with the highest rainfall occurring from September to November and a sec-
ondary rainy season from April to June. The least rain falls from December to March,
and rainfall is also moderately low from July to August (Snow 1976; Lobo-Guerrero
1993). Based on rainfall distribution, the Pacific basins are divided into three zones:
the northern, central, and southern basins. The northern zone, including the water-
sheds of the Atrato, Baudó, and San Juan Rivers (Fig. 14.2), receives on average 5 600 mm
of rainfall annually. The central zone, which includes the watersheds of the Dagua,
Anchicayá, Cajambre, Raposo, Yurumanguí, San Juan de Micay, Iscuandé, Amarales,
Satinga, and Sanguianga Rivers, receives on average 4 100 mm of rainfall annually. The
southern zone, consisting of the drainage basins of the Patía and Mira Rivers, receives
on average 2000 mm of rainfall annually.
The drainage areas, water and sediment transports for the largest Pacific rivers are
synthesised in Table 14.1. Four rivers provide most of the freshwater discharge into the
Pacific. The largest is the San Juan River with a mean discharge of 2550 m3s–1. The
Patía, as gauged at Puente Pusmeo, discharges on average only 328 m3s–1, but the mean
river basin discharge is 1 291 m3s–1 because of the large contribution from the Telembí
River, the last tributary before the delta. The Mira River contributes an average 839 m3s–1,
and the Baudó 782 m3s–1 (Table 14.1). The annual water discharge into the Pacific Ocean
from these four larger rivers and the many smaller Pacific rivers measures 8 020 m3s–
1 or annually 254 km3.
The Patía has the largest drainage basin of the Colombian rivers draining into the
Pacific (23700 km2). From the upper river, sediment loads measure 0.88, 15.39, 13,71
and 8.82 ×106 t yr–1, as gauged at La Fonda, Puente Guascas, Puente Pusmeo, and Los
Nortes, respectively (Table 14.1). Based on daily measurements from 1988 to 1995 by
IDEAM, Instituto de Estudios Ambientales de Colombia (IDEAM 1995), at Los Nortes,
9 km downstream of Puente Pusmeo, and representing an upstream basin area of
14 500 km2, the maximum recorded sediment load was 245.8 ×103 t d–1 in November
1993, and the monthly mean sediment load measured 57.76 ×103 t, corresponding to
an annual sediment load of 21.1 ×106 t yr–1. The sediment yield for the Patía River
ranges from 972 t km–2 yr–1 at Puente Pusmeo to 1 714 t km–2 yr–1 at Puente Guascas,
for the upstream-most portion of the river, the highest yield of any measured river in
Colombia.
The upper portion of the Mira River has an annual sediment load of 0.234 ×106 t yr–1
as gauged at Pipiguay and a sediment yield of 856 tkm–2 yr–1. Because this gauging
station is located 130 km upstream and represents only 4% of the total basin area, the
load is not included in the Pacific budget. Considering the two gauged rivers at their
furthest downstream stations, San Juan and Patía, the measured annual sediment loads
of these rivers into the Pacific Ocean is 30.13 ×106 t yr–1 (Table 14.1). The Atrato River
is a special case. Although the Atrato has its watershed west of the Cordilleras, and
thus have many common characteristics with the Pacific rivers, it discharges into the
Caribbean Sea (Figs. 14.1 and 14.2). Thus, we did not include the Atrato in the Pacific
budgets.
The relation between sediment yield and basin area for the Pacific rivers were de-
termined by log-linear regression of sediment yield on basin area (Restrepo and
Kjerfve 2000a). The analysis included only data for the most downstream gauging
locations on the San Juan, Patía, and Atrato Rivers (Tables 14.1 and 14.2), and in devel-

173CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
oping this regression, we did include the Atrato. Regression of sediment yield on ba-
sin area yielded a decreasing trend for larger basins with a coefficient of determina-
tion, r2= 0.97.
Table 14.1. Drainage basin, annual rainfall, measured water and sediment transports, and calculated
yields for Colombian rivers draining into the Pacific Ocean

J. D. Restrepo · B. Kjerfve174
The sediment load for the non-gauged area of the Pacific coast was obtained from
the regression of sediment yield on basin area, using the gauged data and data for the
San Juan, Patía, and Atrato Rivers (Tables 14.1 and 14.2). The mean sediment yield for
the non-gauged watersheds is 1 827 t km–2 yr–1, occupying a combined area of
36 100 km2, and with a calculated sediment load of 66 ×106 t yr–1. The best estimate of
total sediment load into the Pacific Ocean from both gauged and non-gauged rivers is
96 ×106 t yr–1. These results in a sediment yield estimate of 1 260 t km–2 yr–1, very s imi-
Table 14.2. Drainage basin, annual rainfall, measured water discharge, sediment load, and calculated
yields for the Caribbean Rivers of Colombia

175CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
lar to the yield of 1 200 t km–2 yr–1 proposed by Milliman and Syvitski (1992), based
on extrapolation of data for a single river in Peru.
The sediment yields of the upstream San Juan River at Tadó (1 570 t km–2 yr–1) and
the upstream Patía River at Puente Guascas (1714 t km–2 yr–1) (Table 14.1) are substan-
tially higher than the averages calculated for the entire Pacific coast, and are among
the highest values anywhere in the world. The corresponding drainage areas are
1 661 km2 and 8 900 km2, respectively. Both rivers descend rapidly from the high Cor-
dilleras to the alluvial plain. Over a distance less than 75 km, the San Juan River falls
abruptly from an elevation of 3900 m to 100 m at Tadó (Fig. 14.2), and the tribute ba-
sin descends from elevations between 4 200 and 2500 m in less than 50 km to join the
San Juan River in the upper watershed at an elevation of 90 m. Likewise, the Patía River
descends from its headwaters at 4580 m elevation to 400 m over a distance of 150 km.
Since the San Juan drainage basin as a whole has a greater sediment yield compared
to the Patía drainage basin (Table 14.1), the explanation for the higher yield of
the upstream portions of the Patía implies greater sediment deposition (storage)
on the alluvial plains of the Patía. In the case of the San Juan River, the control exerted
by the Tertiary formations in the middle and lower courses of the river results in a
much narrower alluvial plain as compared to Patía River, and thus less sediment depo-
sition/storage within the drainage basin.
14.2.2
The Caribbean Basins
Caribbean Colombia is principally drained by the Magdalena and Sinú Rivers, and also
receives the Atrato drainage from west of the Cordilleras (Figs. 14.1 and 14.2). The
Magdalena River measures 1612 km and drains a 257 438 km2 basin, which occupies a
major portion of the Colombian Andes. It is the largest fluvial system in Colombia and
originates from headwaters in the Andean Cordillera at an elevation of 3 300 m. The
Sinú River empties into the Morrosquillo Gulf. The Atrato, draining a basin of
35 700 km2, occupies a considerable portion of the Pacific basin, but the river empties
into the Caribbean via the Urabá Gulf (Fig. 14.2).
Analysis of 22 rivers draining into the Caribbean Sea indicates that the combined
water discharge and sediment load are 338 km3yr–1 and 168 ×106 t yr–1, respectively,
corresponding to a sediment yield for the Colombia Caribbean drainage basins of
541 t km–2 yr–1, or approximately half of the yield for the Pacific basins of Colombia
(Table 14.2).
Based on discharge gauging and sediment concentration measurements, the sedi-
ment load of the Atrato River is 11.3 ×106 t yr–1, and the corresponding sediment yield
315 t km–2 yr–1. The sediment yield is comparatively low because of the large size of
the drainage basin and the extensive low-lying Urabá alluvial flood plains, with an area
of 5 500 km2, where significant sediment deposition and storage occur. Besides the
Atrato, several other rivers discharge into the Urabá Gulf (Fig. 14.2). These rivers are
characterised by having small drainage basins and high sediment yields (Table 14.2).
The Sinú River empties into the Golfo de Morrosquillo (Fig. 14.2) and drains an area
of 10 180 km2. Based on monthly data from 1963 to 1993, the annual discharge of the
Sinú is 373 m3s–1. The sediment load is 6 ×106 t yr–1, based on data from 1972 to 1993,
with a sediment yield of 589 t km–2 yr–1 at Montería (Table 14.2).

J. D. Restrepo · B. Kjerfve176
The Magdalena River is the largest river system with a length of 1 500 km. It drains
the Andes Cordillera, which forms the Western, Central, and Eastern Cordilleras. The
drainage basin area measures 257438 km2 and occupies a considerable part of the
Colombian Andes. Daily water discharge measurements from 1975 to 1995 at Calamar
indicate an annual discharge of 7 232 m3s–1. Load measurements during the 21-year
period yielded an annual sediment load of 144 ×106 t yr–1. The calculated sediment
yield for the drainage basin area upstream of Calamar is 559 t km–2 yr–1. The Canal
del Dique (Fig. 14.2) is a 114 km long man-made channel from the Magdalena River at
Calamar to Bahía de Cartagena and was constructed in 1514 by native slaves by order
of Spanish conquistadors. The mean annual water discharge and sediment load through
this channel are currently 299 m3s–1 and 4.8 ×106 t yr–1, respectively (Table 14.2).
The Magdalena River contributes 9% of the total sediment load discharged from
the east coast of South America. The 144 ×106 t yr–1 estimate of sediment load is higher
than the 133 ×106 t yr–1 reported by Marín (1992) but considerably lower than the estimate
by Milliman and Meade (1983) of 220 ×106 t yr–1. Our sediment load estimate sediment
implies a sediment yield of 559 t km–2 yr–1 for the Magdalena, which is more realistic than
the previously reported values of 1000 t km–2 yr–1 (Meybeck 1976, 1988), 900 t km–2 yr–1
(Milliman and Meade 1983), and 920 t km–2 yr–1 (Milliman and Syvitski 1992).
14.3
Dissolved Load
Major natural origins and controls of river-borne materials include atmospheric in-
puts, chemical weathering of mineral, mechanical erosion of rock and soil particles,
and soil leaching. As a result, rivers contain naturally occurring compounds, e.g. major
ions (i.e. Ca2+, Mg2+, and HCO3
–), plant nutrients (e.g. SiO2, NO3
–, NH4
+, and orthophos-
phates), organic compounds (e.g. humic acids and hydrocarbons), and xenobiotic sub-
stances synthesised by humans (Meybeck 2001b).
In the Andes, sedimentary rocks constitute the principal basement lithology, and the
river chemistry agrees with basin geology (Stallard 1980, 1985; Stallard and Edmond 1983).
The concentrations of major dissolved constituents and mass transport rates for ma-
jor Colombian rivers including the larger Magdalena, El Dique, Sinú, Atrato, Mira, an d Pa tía
(Fig. 14.2) are shown in Table 14.3. Estimates are based on averages calculated from
monthly samples from 1990 to 1993 (IDEAM 1995). Ca2+ and Mg2+ are the dominant
ions (Table 14.3), indicating that the water corresponds to the rock-dominated type.
The inorganic carbon concentration was well within the common range of river
pH values, which vary between 6 and 8.2. It is 100% due to atmospheric CO2 and soil
weathering in non-carbonate basins, whereas in carbonate basins 50% comes from the
dissolution of carbonates and other rocks (Meybeck 1996; Knighton 1998). Dissolved
inorganic carbon, present mostly as bicarbonate ions, constitutes almost 50% of the
TDS in the Colombian rivers (Table 14.3). High values of alkalinity seem to be well
explained by high rates of total dissolved solids (TDS) in all Colombian rivers. Regres-
sion of alkalinity concentration (mg l–1) on TDS (mg l–1) yielded a coefficient of de-
termination of 0.98, accentuating the predominance of bicarbonates.
Values of solute concentrations show that the Sinú and Magdalena have the high-
est dissolved solute content followed by the El Dique canal and Patía River (Table 14.3).
The Atrato is by far the most dilute river, four times less mineralised than the larger

177CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
Magdalena and the other Caribbean rivers, as a result of its location in a very humid
environment. The upper and middle sections of the Atrato are located in regions with
very high annual rainfall. The meteorological station at Granja Agrícola Lloró in the
upper Atrato basin at an elevation of 120 m has an annual rainfall rate of 12 717 mm,
based on data from 1952 to 1989 (Eslava 1992). This, to the best of our knowledge, rep-
resents the highest rainfall rate anywhere in South America.
Also, depending on local or regional conditions, natural chemical water composi-
tion can differ by two or three orders of magnitude between basins. As a consequence
of multiple controls on river chemistry including lithology, climate, and topography,
it is inappropriate to refer to any continental or regional averages for comparisons to
local data (Meybeck 1996, 2002). In heterogeneous mountains basins, e.g. in the Co-
lombian Andes, stream and river chemistry are highly variable. More detailed studies
are needed to establish the natural controls on solute concentrations by each rock type.
Chemical weathering of rocks still remains the main source of dissolved substances.
Climate determines where tropical weathering occurs, while tectonics increase ero-
sion rates and dictate the composition of erosion products. In the humid tropics, the
primary factor that distinguishes different weathering regimes is tectonic setting. In
tectonically active areas, easily weathered lithologies are exposed on steep slopes and
weathering rates are lithology dependent (Stallard 1988). Where carbonates and cat-
Table 14.3. Basic hydrochemical data and dissolved solutes of major Caribbean and Pacific rivers of
Colombia for the period 1990–1993. Solute values are expressed as discharge-weighted mean. TSS = total
suspended solids; TDS =total dissolved solids (Source: IDEAM 1995)

J. D. Restrepo · B. Kjerfve178
ions are present, their weathering products dominate the river solution chemistry
(Table 14.3). Furthermore, the presence of unstable and cation-rich minerals in the
suspended load and bedload of rivers draining the Andean basins indicates that rapid
erosion is indeed occurring. Thus, along the western portion of the Pacific basins, high
temperature, humid conditions, high rainfall, and abundant vegetation promote rapid
chemical weathering and high denudation rates (Table 14.1).
The hydrological regime of rivers is a major regulator of their chemical composi-
tion. For each chemical element or TDS value, concentrations and fluxes are discharge
dependant (Meybeck 1996, 2001a,b). The estimates of dissolved materials exported to
the Caribbean and Pacific basins are mainly controlled by water discharge. Thus, the
Magdalena transports 30 ×106 t yr–1 of dissolved materials into the Caribbean
(Table 14.3). The specific transport rate is highest in the Sinú basin, 167 t km–2 yr–1, fol-
lowed by that of the Magdalena (117 t km–2 yr–1). The Atrato, Mira, and Patía Rivers have
values ranging between 31 and 90 t km–2 yr–1.
14.4
Interannual Variability
All South American rivers, independent of size, display a strong seasonal signal of dis-
charge and sediment load variability, typically a factor of 5–10 comparing low monthly
to high monthly discharge (Fig. 14.3). The interannual variability of discharge and
sediment load associated with the ENSO or El Niño-La Niña cycle can be almost equally
great, typically a factor of 2–4, comparing low annual to high annual discharges (Richey
et al. 1986, 1989; Depetris et al. 1996; Vörösmarty et al. 1996). This variability can be
quantified by the southern oscillation index (SOI), which is defined as the difference
in atmospheric sea-level pressure between Tahiti and Darwin (Glantz 1997). The cold
La Niña phase of the SOI is characterised by a positive peak SOI of approximately
+5 hPa, whereas the warm El Niño phase is characterised by a negative peak SOI of
approximately –5 hPa (Fig. 14.3a). The El Niño-La Niña cycle gives rise to a significant
variability in regional rainfall, river discharge, and sediment load. However, the north-
ern and southern portions of the South American continent have a response that is
completely opposite in phase.
El Niño brings about heavy rainfall south of a hypothetical line from Quito, Ecua-
dor, to São Paulo, Brazil. The rivers respond with large increases in both water dis-
charge and sediment load during the Southern Hemisphere’s late summer, when ex-
tensive river flooding impact Paraná and Santa Catarina, Brazil, the delta of the Paraná
River in Argentina, and many other river basins in the south of the continent (Mechoso
and Perez-Iribarren 1992; Probst and Tardy 1989). This causes destructive and costly
flooding of cities, roads, and agricultural fields and brings about much hardship. At
the same time, river basins in South America, north of the front, suffer from drought
conditions and low river discharges, which have negative impacts on the regional ag-
riculture and water resources.
In contrast, during the La Niña phase, the southeast trade winds are well developed,
and the intertropical convergence zone (ITCZ) remains north of its typical position
in the Eastern Pacific. This results in drier than normal conditions in the southern

179CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
Fig. 14.3. Time series plots of mean monthly (thin lines) and low-frequency pass filter with zero
phase (bold lines); a water discharge for San Juan River 1970–1994; b the southern oscillation index
(SOI) (National Oceanic and Atmospheric Administration, NOAA, 1999; data-base on the internet at
http://ftp.ncep.noaa.gov/pub/cpc/wd52dg/data/indices); c water discharge for Magdalena River
1975–1995 (modified from Restrepo and Kjerfve 2000a,b)

J. D. Restrepo · B. Kjerfve180
portion of the South American continent but brings about intense rainfall in the north-
ern parts of the continent (Ropelewski and Halpert 1987). Rivers in Colombia
(Figs. 14.3a and 14.3c) and Venezuela, in particular, experience catastrophic flood con-
ditions, which often have drastic social and economic impacts.
The San Juan River discharge and the smoothed monthly values of the SOI showed
very good coherence for the 25-year period from 1970 to 1994 (Fig. 14.3a). Peak flow
exceeds 5000 m3s–1 during La Niña years and low discharges of 600–1 500 m3s–1 were
observed during El Niño years. Mean annual discharge during El Niño and La Niña
years are 3 625 m3s–1 and 1 490 m3s–1, respectively. Regression analysis of smoothed
SOI on smoothed discharge yielded a coefficient of variation of R2= 0.60, which indi-
cates that variations in the SOI explain 60% of the variability in discharge, with high
values of the SOI corresponding to peak La Niña conditions and peak San Juan dis-
charge. This relationship is similar to the response of Rio Orinoco but contrary to riv-
ers in Perú, Rio Guaiba (Brazil), Rio Paraná (Argentina) (Goniadzki 1999), and other
rivers, which experience significantly higher discharges during the warm El Niño
phase.
In the Magdalena River, water discharge varies significantly interannually. The mean
discharge is 7 200 m3s–1, and the seasonal root mean square (rms) variability is
2 020 m3s–1. The Magdalena discharge at the Calamar station (Fig. 14.2) and the
smoothed monthly values of the SOI show very good coherence for the 21-year period
from 1975 to 1995. Peak flows usually exceed 12 000 m3s–1 during La Niña years, and
low discharges of 2 000–3 000 m3s–1 are observed during El Niño years (Fig. 14.3c).
Mean annual discharges during El Niño and La Niña years are 5 512 m3s–1 and
8 747 m3s–1, respectively.
14.5
The Colombian Rivers and the Global Trend
Rivers with smaller basins have less area to store sediments, and the sediment yield of
smaller basins increases as much as sevenfold for each order of magnitude decrease
in basin area. The result is that many rivers draining smaller basins can have higher
yields than rivers draining larger basins (Milliman 1990; Milliman and Syvitski 1992).
In comparing rivers with small basins in high rainfall areas in both Colombia and
Asia/Oceania (Fig. 14.4), the San Juan and Patía Rivers are similar in terms of water
discharge, sediment load and yields, to the Purari and Fly Rivers in Papua New Guinea.
Average annual rainfall ranges from 2000 mm to 8500 mm in the 33670 km2 catch-
ment of the Purari, which has a mean discharge of 2 360 m3s–1 (Pickup 1983). The Fly
River has a mean discharge of 2 390 m3s–1 and a sediment yield of 1 500 t km–2 yr–1
(Pickup et al. 1981). Although the San Juan drains a basin approximately half as large
as the basin of the Purari (33 670 km2) and far smaller than the 76 000 km2 size of the
Fly, it has greater water discharge. The yield of the Fly, 1500 t km–2 yr–1, is very similar
to the yield of the upper San Juan and Patía Rivers (Table 14.1).
Mountainous rivers with basin areas of ~10 000 km2 in southeast Asia/Oceania have
sediment yields between 140 and 1 700 t km–2 yr–1, and have higher yields by a factor
of 2–3 than rivers draining most other mountainous areas of the world. The Pacific
basins of Colombia have yields between 1 150 and 1 714 t km–2 yr–1, which are very simi-

181CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
lar to rivers draining mountainous terrain and high rainfall areas in South Asia and
Oceania (Fig. 14.4).
The data shown in Tables 14.1–14.3 confirm that the Pacific rivers have smaller drain-
age basins but much higher yields than the Caribbean rivers and those draining east-
ern South America (Amazon, Orinoco, Paraná) (Fig. 14.4). The Patía and San Juan Riv-
ers appear to have the highest sediment yields of any river in South America, and have
many characteristics comparable to the rivers of Papua New Guinea and Taiwan, based
on rainfall, mountainous terrain, small river-basin area, and high sediment transport.
The Magdalena River has the highest sediment yield of the large rivers along the
Caribbean and Atlantic coasts of South America. Its yield is almost three times greater
than the yield of the Amazon, 190 t km–2 yr–1, Orinoco, 150 t km–2 yr–1, or Negro (Ar-
gentina), 140 t km–2 yr–1 (Milliman and Syvitski 1992), and much greater than the yield
of the Paraná, 30 t km–2 yr–1 (Milliman and Syvitski 1992; Goniadzki 1999), Uruguay,
45 t km–2 yr–1, and São Francisco, 10 t km–2 yr–1 (Milliman and Syvitski 1992)
(Table 14.4).
The dissolved load for the Magdalena, 30 ×106 tyr
–1 (Table 14.3), is of the same
magnitude as the Orinoco (30.5 ×106 t y r–1; Depetris and Paolini 1991), ten times lower
than that of the Amazon (259 ×106 t yr–1; Meybeck 1976), and similar to the Parana
River (38.3 ×106 t yr–1; Depetris 1976; Depetris and Paolini 1991) (Table 14.4).
Fig. 14.4. Variation of sediment yield with basin area for several mountainous rivers of Asia/Oceania,
South America (Amazon, Orinoco, and Paraná) and Colombian rivers draining into the Caribbean Sea
(Magdalena, Sinú, and Atrato) and into the Pacific Ocean, San Juan and Patía

J. D. Restrepo · B. Kjerfve182
The major rivers of Colombia fit well into the global river chemistry classification
developed by Gibbs (1970), with Ca2+ and HCO3
– dominating the ionic composition.
Also, values of dissolved solutes are in the range of the most common natural concen-
tration (MCNC) found in most rivers. This classification was proposed by Meybeck
and Helmer (1989) to replace the “average world river,” which is greatly influenced by
a few rivers of extreme concentrations. Thus MCNC is simply the median value of the
distribution of concentrations found in pristine major rivers, weighted by the river
discharge. The ionic natural composition by Colombian rivers with respect to
Ca2+ >Mg
2+ >Na
+>K
+
and HCO3
–>SO
4
2– is similar to the MCNC of other world riv-
ers (cf. Meybeck 1996).
14.6
Environmental Implications
During the past fifty years, the Caribbean rivers and downstream coastal areas have
been under increasing environmental stress. Economic development in Colombia be-
tween the 1970s and 1980s increased demand for river control and utilisation. Ongo-
ing trends in the drainage basins include (1) escalating population densities along the
basins and at the river mouths. The main cities of Colombia, including Bogotá, Medellín,
Cali, and Barranquilla are located in the Magdalena basin (Fig. 14.2). As much as 80%
of the entire population of Colombia lives in the Magdalena watershed, leading to a
demographic density of 54 inhabitants km–2, which is very high when compared to
0.24 inhabitants km–2 in the Amazon basin (Serruya and Pollingher 1984; Depetris and
Paolini 1991); (2) accelerating upland erosion rates due to poor agricultural practices,
Table 14.4. Drainage basin, water discharge, sediment and dissolved loads, calculated yields, and re-
ceiving basin for some rivers of South America (from Depetris 1976; Meybeck 1976; Milliman and Meade
1983; Depetris and Paolini 1991; Milliman and Syvitski 1992; Goniadzki 1999; Restrepo and Kjerfve
2000a,b)

183CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
increasing deforestation and gold mining; and (3) as a result of poor agricultural prac-
tices and deforestation, increasing levels of water pollution. As a result, river-induced
impacts have produced distortion of natural hydrographs, in turn leading to the loss
of critical habitat, biodiversity, and altered material transport (Colciencias-Fen 1989;
HIMAT-INGEOMINAS 1991; Restrepo and Kjerfve 2000b). Although these facts have
been widely recognised, there are until now no quantification of the fluxes for the
Magdalena River.
Sediment load of the Magdalena River has strongly impacted the coastal ecosys-
tems. Since 1954, the government of Colombia has dredged the El Dique Canal, a 114 km
man-made channel from the Magdalena River at Calamar to Cartagena Bay (Fig. 14.2).
Because of increased sedimentation in Cartagena Bay during the 1970s, new canals were
constructed from El Dique to Barbacoas Bay, and since then, the suspended sediment
load in Barbacoas has reached and impacted the El Rosario Islands (Fig. 14.2), a 68 km2
coral reef ecosystem in the Caribbean Sea. Sediment load is responsible for most of
the observed coral reef mortality, with dead coral reef cover reaching 58% (Vernette
1985; INVEMAR 2000b). Also, the suspended sediment load from the Sinú River is
probably responsible for the impact on the largest coral reef on the Colombian Carib-
bean coast, the San Bernardo and Fuerte Islands, a 135 km2 coral reef community north
and south of the Morrosquillo Gulf (Fig. 14.2). Live coral has, in some areas, decreased
25% of the 1995 cover (INVEMAR 2000b).
Water diversion due to the construction of a highway in the Magdalena delta/la-
goon complex, the Ciénaga Grande de Santa Marta, has resulted in hypersalinisation
of mangrove soils and the consequent die-off of almost 270 km2 of mangrove forests
during the past 39 years. Between 1956 and 1995, 66% of the original mangrove forest
died (Botero 1990; Cardona and Botero 1998). Recent estimates indicate that for the
whole Magdalena lagoon/delta complex and associated coastal zones, the mangrove
area has been reduced from 62000 ha in 1991 to 52 478 ha in 1996, almost 2000 ha yr–1
(INVEMAR 2000b). In addition, freshwater input from the Magdalena River to the
lagoon was also diverted for irrigation purposes and interrupted by dikes built along the
delta distributaries to prevent flooding of farmlands. The changes in the hydrological
regime have also caused water quality changes in the lagoons and canals, resulting in
low dissolved oxygen concentration, fish kills, and eutrophication (Botero 2000).
Fluvial geochemistry and material fluxes have already been much altered on the
global scale by agriculture, deforestation, mining, urbanisation, industrialisation, ir-
rigation, and damming. The continental aquatic systems are now affected by hypoxia,
eutrophication, salinisation, and contamination by nitrate, metals, and persistent or-
ganic pollutants. Phosphate (PO4
3–) and nitrate (NO3
–) increases are observed in most
rivers exposed to human pressure (Meybeck 2001b). Their sources are multiple. Since
the 1950s, the use of nitrogen and of phosphorous, both as fertilizers, and in the food,
detergent, and other industries, have resulted in a rapid increase of fluvial N and P
fluxes, now exceeding the pristine levels by a factor of ten in some world rivers
(Meybeck 2002).
In Colombia, pristine fluvial systems like those draining the Pacific basins have
much less PO4
3– and NO3
– loads when compared to the Caribbean rivers (Table 14.5).
The Magdalena and Atrato Rivers are the Colombian systems that contribute by far
the highest P and N fluxes to the sea, with total phosphate and nitrate fluxes up to
186 ×103 tyr
–1 and 47 ×103 tyr
–1, respectively (Table 14.5). Many causes are responsible

J. D. Restrepo · B. Kjerfve184
for these high nutrient loads, including massive sewage collection in cities and towns
for NH4
+ and PO4
3–, mainly in the Magdalena basin, and also due to fertilization of ba-
nana plantations in the lower course of the Atrato River.
Magdalena is the major collector of municipal and industrial waste waters in Co-
lombia. Urban, agricultural, mining, and industrial waste inputs from the Magdalena ba-
sin have aggravated the conditions of the Ciénaga Grande lagoon and coastal ecosystems.
Biodiversity has been reported to be considerably lower in the area affected by man-
grove mortality as well as in the coastal zone (Botero 2000; INVEMAR 2000a). Declining
fisheries from 63 700 tons in 1978 to 7 850 tons in 1998, an approximate decline of eight
times in less than 20 years, is a strong declination for any living resource, and indicates
low environmental and water quality conditions as well as the absence of policies and
management (Beltrán et al. 2000). The fluvial inputs of the Magdalena River into the
Caribbean have great environmental and economic impacts on the coastal ecosystems.
14.7
Conclusions
The results indicate that the sediment yield in the smaller Pacific rivers, San Juan and
Patía (Figs. 14.1 and 14.2), is significantly higher than for larger river basins draining
into the Caribbean and Atlantic Oceans (Table 14.4). However, the Magdalena, the larg-
Table 14.5. Nutrient fluxes of phosphate (PO4
3–) and nitrate (NO3
–) in pristine Pacific rivers and non-
pristine fluvial systems of the Caribbean basins of Colombia. Nutrient values are based on averages
calculated from monthly samples covering the three-year period 1998–2000 (Source: INVEMAR 2000a,
2001; Restrepo and Kjerfve 2000a)

185CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
est river discharging directly into the Caribbean Sea, has the highest sediment yield
of any medium-sized or large river along the entire east coast of South America. This
is consistent with the global trend of sediment yield decreasing for larger basins.
The data also confirm that the San Juan River appears to have the highest yield of
any documented river in South America and has many characteristics comparable to
rivers in Papua New Guinea and Taiwan, based on drainage basin characteristics, rain-
fall, relief, area, and sediment load. The rivers in Colombia exhibit significant discharge
variability as a result of ENSO.
Many Colombian rivers, including the larger Magdalena, are affected by deforesta-
tion and rapid changes in land use, thus accelerating the transfer of particulate and
dissolved organic and inorganic matter from the river basins to the sea. Due to the
magnitude of fluvial fluxes to the oceans from the Colombian rivers (Tables 14.4 and
14.5), the fluctuations of dissolved and suspended loads need to be monitored for a
period of at least ten years, in order to be able to quantify the influences of man’s ac-
tivities and assess global climate.
References
Beltrán IC, Estrada M, Valderrama M (2000) Plan de Ordenación: manejo y aprovechamiento sostenible
pesquero y acuicola de la Cuenca del Río Grande de la Magdalena. INPA, Bogotá, Colombia (Techni-
cal Report, pp 1–37)
Botero L (1990) Massive mangrove mortality on the Caribbean Coast of Colombia. Vida Silvestre
Neotropical 2:77–78
Botero L (2000) Restoration trajectory of structural and functional processes in a tropical mangrove-
estuarine ecosystem (Ciénaga Grande de Santa Marta) in Caribbean Colombia. In: Lacerda LD,
Kremer HH, Salomons W, Kjerfve B, Crossland JM (eds) South American river catchment/coastal
zone interaction and human dimensions. LOICZ Report 37:39–44
Cardona P, Botero L (1998) Soil characteristics and vegetation structure in a heavily deteriorated man-
grove forest in the Caribbean Coast of Colombia. Biotropica 30:24–34
COLCIENCIAS-FEN (1989) Perfil Ambiental Colombia, 2nd edn. Escala Editores, Bogotá
Depetris PJ (1976) Hydrochemestry of the Paraná River. Limnol Oceanogr 21:736–739
Depetris PJ, Paolini JE (1991) Biogeochemical aspects of South American rivers: the Paraná and Orinoco.
In: Degens TE, Kempe S, Richey JE (eds) Biogeochemistry of major world rivers. Wiley, Chichester,
pp 105–125
Depetris PJ, Kempe S, Latif M, Mook WG (1996) ENSO-Controlled flooding in the Paraná river
(1904–1991). Naturwissenschaften 83:127–129
Eslava J (1992) La precipitación en la región del Pacífico colombiano (Lloró: el sitio más lluvioso del
mundo). Revista Zenit 3:47–71
Gibbs RJ (1970) Mechanisms controlling world water chemistry. Science 180:1088–1090
Glantz MH (1997) Currents of change, El Niño’s impact on climate and society. Cambridge University
Press, Cambridge
Goniadzki D (1999) Hydrological warning and information system for the Plata basin. Instituto Nacional
del Agua y del Ambiente, Buenos Aires
HIMAT-INGEOMINAS (1991) Estudio de la contaminación del Río Magdalena por metales traza, su relación
con parámetros hidrobiológicos, fisicoquímicos y su incidencia en la salud humana. Informe, B ogotá
IDEAM (1995) Estadísticas hidrológicas de Colombia (1990–1993). Tomos I–II. Diego Samper Ediciones,
Bogotá
INVEMAR (2000a) Proyecto de monitoreo de las condiciones ambientales y de los cambios estructurales
y funcionales de las comunidades vegetales y los recursos pesqueros durante la rehabilitación de la
Ciénaga Grande de Santa Marta: un enfoque de manejo adaptativo. Fase I: 1999–2000. INVEMAR,
Santa Marta, Colombia (Technical Report)
INVEMAR (2000b) Informe anual sobre el estado de los recursos marinos y costeros. Programa SINAM-
Unidad Coordinadora de Informacion. INVEMAR, Santa Marta, Colombia (Technical Report)
INVEMAR (2001) Diagnóstico y evaluación de la calidad ambiental marina en el Caribe y Pacífico
Colombiano. Red de Vigilancia para la Conservación y Protección de las Aguas Marinas y Costeras
de Colombia. Informe Técnico Final, Convenio 1060/2000 MMA-FONAM-INVEMAR (compact disk)

J. D. Restrepo · B. Kjerfve186
Kellog J, Mohriak WU (2001) The tectonic and geological environment of coastal South America. In:
Seeliger U, Kjerfve B (eds) Coastal marine ecosystems of Latin America. Springer-Verlag, Berlin
Heilderberg, pp 2–16
Knighton D (1988) Fluvial forms and processes. John Wiley and Sons, New York
Lobo-Guerrero A (1993) Hidrología e hidrogeología. In: Leyva P (ed) Colombia Pacifíco – Tomo I. Fondo
para la Protección del Medio Ambiente AJosé Celestino Mutis- FEN, Bogotá, pp 120–134
Marín R (1992) Estadísticas sobre el recurso agua en Colombia, 2ed edn. Ministerio de Agricultura,
Instituto Colombiano de Hidrología, Metereología y Adecuación de Tierras, Bogotá
Mechoso CR, Perez-Iribarren G (1992) Streamflow in the southeastern South America and the south-
ern oscillation. J Climate 5:1535–1539
Meybeck M (1976) Total mineral transport by world rivers. Hydrol Sci Bull 21:265–284
Meybeck M (1988) How to establish and use world budgets of riverine materials. In: Lerman A, Meybe ck
M (eds) Physical and chemical weathering in geochemical cycles. NATO ASI Series C in Mathemati-
cal and Physical Sciences 51:247–272
Meybeck M (1996) River water quality: global ranges, time and space variabilities, proposal for some
redefinitions. Verh Internat Verein Limnol 26:81–96
Meybeck M (2001a) River basins under anthropocene conditions. In: Bodungen B, Turner RK (eds) Sci-
ence and integrated coastal management. Dahlem University Press, pp 275–294
Meybeck M (2001b) Global alteration of riverine geo chemestry under human pressure. In: Ehlers E, Kraft
S (eds) Understanding the Earth system: compartments, processes and interactions. Springer,
Heidelberg, pp 97–113
Meybeck M (2002) Water quality. In: Goudie A, Cuff DJ (eds) Encyclopedia of global change, environ-
mental change and human society. Oxford University Press, Oxford, pp 503–515
Meybeck M, Helmer R (1989) The quality of rivers: from pristine stage to global pollution. Global Plan-
etary Change 1:283–309
Milliman JD (1990) Fluvial sediment in coastal seas: flux and fate. Nature and Resources (Unesco) 26:12–
22
Milliman JD, Meade RH (1983) World-wide delivery of river sediment to the oceans. J Geol 91:1–21
Milliman JD, Syvitski PM (1992) Geomorphic/tectonic control of sediment transport to the ocean: the
importance of small mountainous rivers. J Geol 100:525–544
NOAA (1999) Tahiti-Darwin sea level pressure anomaly. (http://ftp.ncep.noaa.gov/pub/cpc/wd52dg/
data/indices)
Pickup G (1983) Sedimentation processes in the Purari River upstream of the delta. In: Petr T (ed) The
Purari: tropical environment of a high rainfall basin. Dr. W. Junk Publishers, Boston, pp 205–226
Pickup G, Higgins RJ, Warner RF (1981) Erosion and sediment yield in Fly River drainage basins, Papua
New Guinea. Int Assoc Hydrol Sci Pub 132:438–456
Probst JL, Tardy Y (1989) Global runoff fluctuations during the last 80 years in relation to world tem-
perature change. Am J Sci 289:267–285
Restrepo JD, Kjerfve B (2000a) Water discharge and sediment load from the western slopes of the Co-
lombian Andes with focus on Rio San Juan. J Geol 108:17–33
Restrepo JD, Kjerfve B (2000b). Magdalena River: interannual variability (1975–1995) and revised water
discharge and sediment load estimates. J Hydrol 235:137–149
Restrepo JD, Kjerfve B (2002) The San Juan delta, Colombia: tides, circulation, and salt dispersion. Con-
tinental Shelf Res 22:1249–1267
Richey JE, Meade RH, Salati E, Devol AH, Nordin CF, dos Santos U (1986) Water discharge and suspended
sediment concentrations in the Amazon River: 1982–1984. Water Resources Res 22:756–764
Richey JE, Nobre C, Deser C (1989) Amazon River discharge and climate variability: 1903 to 1985. Sci-
ence 246:101–103
Richey JE, Victoria RL, Forsberg BR (1991) The biochemestry of a major river: the Amazon case study.
In: Degens TE, Kempe S, Richey JE (eds) Biogeochemistry of major world rivers. Wiley, Chichester,
pp 57–104
Ropelewski CF, Halpert MS (1987) Global and regional scales precipitation associated with El Niño-South-
ern Oscillation. Monthly Weather Review 115:1606–1626
Serruya C, Pollingher U (1984) Lakes of the warm belt. Cambridge University Press, Cambridge
Snow JW (1976) The climate of northern South America: Colombia. In: Schwerdtfeger W (ed) Climates
of South and Central America. Elsevier Scientific Publishing Company, Amsterdam, pp 358–379
Stallard RF (1980) Major element geochemestry of the Amazon River system. Massachusetts Institute
of Technology (Joint program in oceanography and ocean ingineering, WHOI-80-29)
Stallard RF (1985) River chemistry, geology, geomorphology, and soils in the Amazon and Orinoco ba-
sins. In: Drever JI (ed) The chemistry of weathering. D. Reidel Pub. Co., Dordrecht (Proceedings of a
NATO Advanced Research Workshop)

187CHAPTER 14 · The Pacific and Caribbean Rivers of Colombia
Stallard RF (1988) Weathering and erosion in the humid tropics. In: Lerman A, Meybeck M (eds) Physi-
cal and chemical weathering in geochemical cycles. NATO ASI Series C in Mathematical and Physi-
cal Sciences 51:225–246
Stallard RF, Edmond JM (1983) Geochemestry of the Amazon. 2. The influence of geology and weather-
ing environment on the dissolved load. J Geophys Res 88:9671–9688
Ver nette G (1985) La plateforme continentale caraibe de Colombie (du débouché du Magdalena au Golfe
de Morrosquillo). Importance du diapirisme argileux sur la morphologie et la sédimenattion. PhD
dissertation, Université Bordeaux I, Bordeaux
Vörösmarty CJ, Willmott CJ, Choudhury BJ, Schloss AL, Stearns TK, Robeson SM, Dorman TJ (1996)
Analyzing the discharge regime of a large tropical river through remote sensing, ground-based cli-
matic data, and modelling. Water Resources Res 32:3137–3150