SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / s c i t o t e n v

Changing climate and endangered high mountain ecosystems

in Colombia

Daniel Ruiz⁎, Hernán Alonso Moreno, María Elena Gutiérrez, Paula Andrea Zapata
Grupo de Investigación ‘Gestión del Ambiente para el Bienestar Social—GABiS', Escuela de Ingeniería de Antioquia, Calle 25Sur No. 42-73,
Envigado, Antioquia, Colombia

AR TIC LE I N FO ABS TR ACT

Article history: High mountain ecosystems are among the most sensitive environments to changes in
Received 20 July 2007 climatic conditions occurring on global, regional and local scales. The article describes the
Received in revised form changing conditions observed over recent years in the high mountain basin of the Claro
15 February 2008 River, on the west flank of the Colombian Andean Central mountain range. Local ground
Accepted 24 February 2008 truth data gathered at 4150m, regional data available at nearby weather stations, and
Available online 23 April 2008 satellite info were used to analyze changes in the mean and the variance, and significant
trends in climatic time series. Records included minimum, mean and maximum
Keywords: temperatures, relative humidity, rainfall, sunshine, and cloud characteristics. In high
High mountain ecosystems levels, minimum and maximum temperatures during the coldest days increased at a rate of
Páramos about 0.6°C/decade, whereas maximum temperatures during the warmest days increased at
Glaciers a rate of about 1.3°C/decade. Rates of increase in maximum, mean and minimum diurnal
Global change temperature range reached 0.6, 0.7, and 0.5°C/decade. Maximum, mean and minimum
relative humidity records showed reductions of about 1.8, 3.9 and 6.6%/decade. The total
number of sunny days per month increased in almost 2.1 days. The headwaters exhibited no
changes in rainfall totals, but evidenced an increased occurrence of unusually heavy rainfall
events. Reductions in the amount of all cloud types over the area reached 1.9%/decade. In low
levels changes in mean monthly temperatures and monthly rainfall totals exceeded + 0.2°C
and − 4% per decade, respectively. These striking changes might have contributed to the
retreat of glacier icecaps and to the disappearance of high altitude water bodies, as well as to
the occurrence and rapid spread of natural and man-induced forest fires. Significant
reductions in water supply, important disruptions of the integrity of high mountain
ecosystems, and dramatic losses of biodiversity are now a steady menu of the severe climatic
conditions experienced by these fragile tropical environments.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction disproportionately by the poor, who are the most vulnerable

The well-being of human societies is based on the sustained
delivery of fundamental ecosystem services, such as (among
others) ‘the regulation of the quality and quantity of water
supply, and the security in the face of future environmental
change’ (Díaz et al., 2006). Regrettably, the consequences of the
loss of these important goods and services ‘will be felt
to such environmental problems’ (World Resources Institute,
2005). Colombia is a good example. The United Nations
Framework Convention on Climate Change received the first
Colombian official communication in 2002 (NC1-2002).
The future climate scenario suggests a nationwide increase
of 1–2°C in mean annual temperatures and variations in
annual rainfall totals of about + 15% by 2050. The NC1-2002

⁎ Corresponding author. Tel.: +57 4 339 3200; fax: +57 4 331 7851.
E-mail address: pfcarlos@eia.edu.co (D. Ruiz).

0048-9697/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2008.02.038
 S CIE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32 123

identified the Colombian high mountain ecosystems (HME) as
one of the areas of primary concern (World Bank Group, 2006).
Colombia is host to the largest stretch of páramos life zones
in the planet (WBG, 2006). These are exceptional desolate
regions located only in the Tropics between the high
mountain Andean forests (the so-called bosque montano; ca.
2000–3500m) and the areas of ‘permanent’ snow (ca. N 4500m).
Their climatic conditions are characterized by average tem-
peratures below 10°C, strong diurnal temperature range,
cloudy skies, foggy days, high UV radiation amounts, strong
winds, and light rain. These fragile ecosystems have unique
endemic flora and serve as important sources and reservoirs
of water. Páramos habitats and towering snowcapped peaks
constitute the HME. It is expected that ca. 56% of Colombian
páramos are going to be seriously affected by increases in
temperature by 2050 (WBG, 2006). As páramos ecosystems are
only endemic in high elevation regions (no way up) and
mountain species exhibit a weak ability to adapt to changing
climatic conditions, climate change will undoubtedly result in
a dramatic loss of biodiversity.
To further complicate matters glaciers in the areas of
‘permanent’ snow (above 4500m) are also expected to be
affected by increases in temperature. In the NC1-2002, the
Colombian Institute of Hydrology, Meteorology and Environ-
mental Studies (IDEAM) reported that the loss of the total
Colombian glacier area had reached almost 80% since 1850.
Sadly, it is expected that 78% of the remaining Colombian
glaciers will be seriously affected by increases in temperature
by 2050 (WBG, 2006). The icecaps of these glaciers ‘feed’ high
altitude water bodies and permanent water-reservoir habitats
(the so-called turberas) that contribute to the headwaters of
several rivers currently used by lowland populations to satisfy
the water demand (a significant proportion of Colombian
communities depend directly on high mountains watersheds).
Turberas are mainly colonized by mountain species with
narrow habitat tolerance and low dispersal capacity, and are
likely to be at high risk from the environmental effects of
climate change. Hence, one of the major consequences of the
global environmental problem could be the loss of environ-
mental goods and services provided by these habitats,
especially the water supply and basin regulation (Gutierrez
et al., 2006; WBG, 2006). Unfortunately, dramatic land use
changes (i.e. clearing-off of high mountain Andean forests and
páramos) caused by extensive agriculture and livestock grazing
are also threatening the existence of these fragile reservoirs
and ecosystems, and could be producing severe local climatic
anomalies. As the world is already facing the consequences of
a warmer climate, adaptation strategies are urgently needed
to maintain considerable levels of environmental goods and
services in high mountain ecosystems.

Fig. 1 – General location of the El Ruiz–Tolima Volcanic Massif. The topography is represented using Digital Elevation Models for
30″-arc (924 m × 924 m, left) and 90 m × 90 m spatial resolutions (right). The latter is based on the United States Geological Survey
(USGS)—National Aeronautics and Space Administration (NASA) Digital Terrain Model and was processed using the Geographic
Information System HidroSig Java (Copyright: Postgrado en Aprovechamiento de Recursos Hidráulicos, Escuela de Geociencias
y Medio Ambiente, Universidad Nacional de Colombia Sede Medellín).
124 SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32

In the international arena numerous studies have focused
on the impacts that changes in atmospheric circulation,
temperature, precipitation, water vapor content, and cloud
cover could have on, particularly, mountain glaciers and
consequently on discharge rates and timing in highland rivers
(Kaser and Georges, 1997; Wagnon et al., 1999; Vuille et al.,
2003). Due to their nature, such studies were focused on
regional scales. Vuille et al. (2003), for instance, hypothesized
that changes in temperature and humidity were the primary
cause for the observed retreat of Tropical Andean glaciers
during the second half of the 20th century. Interestingly, these
authors found that near-surface temperatures increased
significantly throughout most of the Tropical Andes, varying
markedly between the Eastern and Western Andean slopes,
with a much larger temperature increase to the west.
In Colombia several studies have been conducted to
determine the most important variables forcing the retreat of
some glaciers (e.g. Euscategui and Ceballos, 1999; Euscategui,
2002) and to quantify the water balance in high mountain
ecosystems (e.g. Diaz-Granados et al., 2005). Euscategui and
Ceballos (1999) and Euscategui (2002) focused their analyses on
the Nevado de Santa Isabel, located on the Colombian Central
Cordillera, and supported their study on some evidences from
the Sierra Nevada del Cocuy, a major glacier located on the
Eastern Andean Cordillera. The authors found ‘close relation-
ships’ between local weather conditions and glacier reces-
sions, and reported a ‘strong incidence’ of El Nino and La Nina
extreme events on glacier retreat. Diaz-Granados et al. (2005)
focused on the hydrologic modeling of the upper Blanco River
basin, located on the Chingaza páramo, considered the major
source of water for Bogotá, the Colombian capital city. These
researchers merged a horizontal precipitation model and a
Geographic Information System to simulate the water supply
in the Blanco River basin, and to predict the potential impacts

Fig. 2 – High-resolution digital terrain model of the Claro River basin (top left) and spatial distributions of mean annual
temperature (top right), annual minimum temperature (bottom left), and mean annual dew point (bottom right). The minimum
and maximum altitudes of the Claro River basin are approximately 1800 and 5200 m. The mean and minimum annual
temperatures range from 18.5 to −2.5 °C and from 12.5 to −6.5 °C, respectively. The mean annual dew point varies from 14.5 to
almost −4.5 °C. The atmospheric pressure ranges from 820 to 560 mb. The potential and actual evapotranspiration rates vary
from 3.2 to 1.2 mm/day and from 2.2 to 1.2 mm/day, respectively. The minimum, actual and saturation vapor pressures range
from 15 to 4 hPa, from 17 to almost 6 hPa, and from 21 to 6 hPa, respectively. Finally, the mean annual relative humidity varies
from 80 to 94%.
 S CIE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32 125

that changes in vegetative cover and land use could have on
water balance. These studies, however, have not been
extensively reviewed by the international community, despite
the importance of their subjects.
All these reported evidences suggest that assessments of
changes in local physical variables are extremely important
for an improved understanding of the mechanisms and
impacts of climate change, and for calibration and verification
of general and regional circulation models. In order to under-
stand evidences of changing climatic conditions and explore
their potential implications on the integrity of high mountain
ecosystems, we aim to analyze several historical time series
(ground truth data and satellite info) to detect significant
changes in the mean and/or the variance, and/or significant
trends in various climatic variables. Analyses focus on a
specific high elevation region located on the west slope of the
Colombian Andean Central Mountain Range (ACMR).
2. Study area and data
2.1. Study area
One of the most important high mountain ecosystems is
located on the ACMR in the ‘Los Nevados’ Natural Park, on the
El Ruiz–Tolima Volcanic Massif (see Fig. 1). Los Nevados
protected area is distant 140km west of Bogotá, has an exten-
sion of about 58,000ha, and has three ice-capped mountains
(the El Ruiz, the Santa Isabel, and the El Tolima) and two high
snow-covered mountains with ephemeral snowfields (the El
Quindio and the El Cisne). These high peaks have summit
elevations ranging from 4900 to 5321m above sea level. An
important watershed of the ‘Los Nevados’ Natural Park is the
high mountain basin of the Claro River, which is located on the
West flank of the Central Cordillera and is currently fed by
numerous mountain torrents originating high in the snow-
fields of the El Ruiz and the Santa Isabel (see Fig. 2). This article
deals with the analysis of possible changing climatic condi-
tions observed over recent years in the Claro River basin.
2.2. Data studied
2.2.1. Regional data
Several cloud characteristics, such as cloud amount, top
pressure, top temperature, and optical thickness were ana-
lyzed for multiple clouds (all types, high-level, middle-level,
low-level, and deep convective clouds) occurring over the
Central Andean region. Datasets are available for the grid
points ID 3556 (03°45′N; 76°15′W), ID 3699 (06°15′N; 76°47′W),
and ID 3700 (06°15′N; 74°16′W) and for the period from July,
1983 through August, 2001 (Rossow et al., 1996). Mean monthly
temperatures and total monthly rainfall records from seven
nearby weather stations of the Global Historical Climate
Network (GHCN) located on both flanks of the Andean Central
Mountain Range were also processed (see Table 1; source:
Baker et al., 1995). All these data were downloaded from the
Data Library of the International Research Institute for Climate
and Society, IRI (http://iri.columbia.edu/).
2.2.2. Topography and climatic conditions
A detailed high-resolution digital terrain model of the Claro
River basin was created using cartography on a 1:25,000 scale,

Table 1 – Historical time series (mean monthly temperatures – top – and monthly rainfall totals – bottom –) available from the
Global Historical Climate Network
WMO station Monthly temperatures (°C) Trend in the
mean
ID Name Latitude Longitude Elevation (m) Available period Mean SD Min Max (°C/decade)

8014902 Chinchina 5.0N 75.6W 1360 Jan/1951–Dec/1970 18.9 0.6 17.4 20.5 +0.2
8014904 Chinchina 5.0N 75.6W 1360 Jan/1951–Dec/1970 20.6 0.6 19.2 22.5 +0.4
8014911 Chinchina 5.0N 75.6W 1310 Jan/1971–Dec/1980 21.0 0.8 19.3 23.6 (Only one decade)

WMO station Monthly rainfall total (mm) Annual Trend in

total the
ID Name Latitude Longitude Elevation Available Mean SD Min Max (mm) mean
(m) period (%/
decade)

8014902 Chinchina 5.0N 75.6W 1360 Jan/1951– 212.3 92.5 26.6 466.0 2,540 −6.0
Dec/1970
8014904 Chinchina 5.0N 75.6W 1360 Apr/1950– 192.5 107.1 0.0 608.0 2,280 −6.0
Dec/1979
8014911 Chinchina 5.0N 75.6W 1310 Jan/1971– 213.4 91.2 9.0 505.9 2,560 (Only one
Dec/1980 decade)
8021000 Pereira 4.8N 75.8W 1338 Jan/1951– 176.8 89.7 5.0 496.0 2,090
Matecana Dec/1993
8021100 Armenia El 4.5N 75.7W 1204 Jan/1950– 166.4 101.6 0.0 455.0 1,970
Eden Mar/1989
8021101 El Paso 4.5N 75.6W 3264 Jan/1950– 152.4 99.8 1.0 572.0 1,810 −4.0
Dec/1979
8021400 Ibague 4.4N 75.1W 928 Jan/1955– 157.3 88.4 1.0 494.0 1,900 −6.0
Perales Dec/1993
126 SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32

available at the Colombian Geographical Institute. The spatial
distributions of the mean and minimum annual tempera-
tures, mean annual dew point (see Fig. 2), and atmospheric
pressure (not shown) were estimated using regression func-
tions on altitude above sea level, according to the equations
discussed by Poveda et al. (2001) for the Andean region. The
potential and actual evapotranspiration rates were estimated
through the equations proposed by the National Coffee
Research Centre, CENICAFE (Chaves and Jaramillo, 1998),
regionalized for the Colombian Central Andes. Finally, the
spatial distributions of the actual vapor pressure, saturation
vapor pressure, mixing ratio, saturation mixing ratio, and
relative humidity were estimated through the Clausius–
Clapeyron equation, adapted to the Andean region.
2.2.3. Local climate data
Total daily sunshine [h], total daily rainfall [mm], daily mini-
mum temperatures [°C], daily maximum temperatures [°C],
and mean daily relative humidity values [%] from a nearby
weather station (ID 2615515 Las Brisas, located at 04°56′N,
75°21′W and altitude 4150m) were also processed. Data per-
taining to the window period comprising from January, 1981 to
December, 2005 were obtained for the analyses. Daily sun-
shine allowed the estimation of the total monthly sunshine,
the total number of cloudy and sunny days, the maximum,
mean and minimum daily sunshine values, and the standard
deviation of daily sunshine records. Daily rainfall totals were
used to estimate the total annual and monthly rainfall values,
the total number of dry days per month and per year, and the
maximum daily rainfall records. Daily minimum and max-
imum temperatures were used to estimate the minimum and
maximum values during the coldest and warmest days, as
well as the standard deviation of daily temperature records.
Daily minimum and maximum temperatures were also used
to estimate the maximum, mean and minimum monthly
diurnal temperature range, as well as the standard deviation
of such climatic variable. Daily relative humidity was used to
calculate the maximum, mean and minimum relative humid-
ity values, and the standard deviation of daily records. De-
tailed information of available and missing periods, as well as
the mean, standard deviation, minimum and maximum va-
lues of each climatic variable is presented in Table 2.
3. Results and discussion
3.1. Observed changes
The total amounts of all cloud types, high-level, middle-level,
low-level, and deep convective clouds occurring over the pixel
represented by the grid point ID 3556, reached average values
of 80.6, 30.9, 34.8, 5.5, and 5.3%, respectively. The reduction in

Table 2 – Historical time series available from Las Brisas local weather station
Climatic variable Available N Total sample Mean Standard Min Max
period size deviation

Total monthly sunshine [h] Jan/1982–Dec/ 288 202 71.9 30.0 17.3 165.8
Total days per month null sunshine [number] 2005 4.9 3.4 0.0 17.0
Total monthly sunny days [number] 24.5 3.8 13.0 31.0
Daily maximum sunshine [h] 8.1 2.1 3.1 11.7
Daily mean sunshine [h] 2.5 1.0 0.6 5.5
Daily minimum sunshine [h] 0.0 0.1 0.0 0.6
Standard deviation daily sunshine [h] 2.3 0.7 0.8 4.0
Total monthly rainfall [mm] Jan/1981–Dec/ 276 239 120.1 61.7 6.5 326.9
Total dry days [number] 2003 218 10.9 5.4 0.0 29.0
Maximum daily rainfall [mm] 21.7 11.7 1.3 99.0
Total annual rainfall [mm] 1981–2003 23 20 1417 231.4 1083 1806
Total annual dry days [number] 20 118 30.6 37 173
Minimum monthly temperatures [°C] — warmest Jan/1981–Dec/ 276 186 3.0 0.9 1.4 7.0
days 2003
Minimum monthly temperatures [°C] 1.2 0.6 − 0.6 3.1
Minimum monthly temperatures [°C] — coldest − 1.0 2.0 − 10.0 1.4
days
Standard deviation minimum temperatures [°C] 1.0 0.4 0.5 3.1
Maximum monthly temperatures [°C] — warmest Jan/1981–Dec/ 276 144 11.3 1.6 8.0 20.0
days 2003
Maximum monthly temperatures [°C] 8.4 1.0 6.0 10.6
Maximum monthly temperatures [°C] — coldest 5.5 1.1 0.0 9.0
days
Standard deviation maximum temperatures [°C] 1.6 0.3 0.9 2.6
Maximum monthly relative humidity [%] Jan/1981–Dec/ 276 184 98.2 2.0 90.0 100.0
Mean monthly relative humidity [%] 2003 92.0 3.9 79.4 98.1
Minimum monthly relative humidity [%] 76.9 8.6 55.0 95.0
Standard deviation relative humidity [%] 5.4 2.0 1.1 10.6
Maximum monthly diurnal temperature range [°C] Jan/1981–Dec/ 276 143 11.2 2.2 7.6 21.0
Monthly diurnal temperature range [°C] 2003 7.2 1.0 4.8 9.6
Minimum monthly diurnal temperature range [°C] 3.6 1.0 0.8 6.6
Standard deviation diurnal temperature range [°C] 2.0 0.5 1.0 4.3
 S CIE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32
Fig. 3 – Historical time series of monthly (gray solid lines) and 3-month moving average (black solid lines) sunshine, temperature and relative humidity from the local weather
station ID 2615515 Las Brisas for the window period comprising from January, 1980 to December, 2005. Top left panel: total number of sunny days per month (TSD) along with
maximum monthly temperatures – average – (ATmax) and maximum temperatures during the warmest (MTmax) and coldest days (mTmax). Top right panel: total number of
sunny days per month (TSD) along with minimum monthly temperatures – average – (ATmin) and minimum temperatures during the warmest (MTmin) and coldest days
(mTmin). Bottom left panel: maximum (MRH), mean (ARH) and minimum (mRH) monthly relative humidity values. Bottom right panel: maximum (MDTR), mean (ADTR) and
minimum (mDTR) monthly diurnal temperature range.

127
128 SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32
the amount of such clouds reached 3.94, 1.95, 1.54, 0.76, and
1.71%, respectively, over the 218-month observing period.
Changes or trends in the time series of top temperatures and
top pressures are not present during the analyzed period.
Trends in the mean of the historical time series of monthly
temperatures and monthly rainfall amounts are observed in
some of the available regional records. Table 1 summarizes
the results for the GHCN weather stations.
Changes in the mean and the variance, as well as
significant trends in some of the historical time series of
sunshine, rainfall, maximum and minimum temperatures,
and relative humidity were observed at the local weather sta-
tion over the period from January, 1981 to December, 2005.
Fig. 3 depicts some of the major findings of the exploratory
analyses. Neither the total monthly sunshine nor the daily
mean and maximum sunshine historical time series exhibited
appreciable trends. However, the total number of days per
month of null sunshine (foggy days) slightly decreased during
the period from January, 1982 through December, 2005, and
hence, the total number of sunny days per month (TSD)
increased by 2.1days over this period. The historical time
series of total annual/monthly rainfall and total number of dry
days per year/month, suggest that no changes in the mean
took place during the period from January, 1981 through
December, 2003. Nevertheless, the maximum daily precipita-
tion values exhibit an increased occurrence of unusually
heavy rainfall events, particularly due to two unprecedented
events of 99 and 85mm/day that happened in January, 1996
and January, 1998, respectively.
The minimum monthly temperatures during the warmest
days (MTmin) decreased from 3.3 to 2.7°C over the period from
January, 1981 through December, 2003. Conversely, the mini-
mum monthly temperatures during the coldest days (mTmin)
increased dramatically from almost − 2.0°C to − 0.3°C over
the same 23-year period (in the 1980s and the early 1990s,
the minimum daily temperatures used to reach values of
almost − 8.0 and − 10.0°C, but in recent years such tempera-
tures have reached only values as cold as − 3.0°C). As a
consequence of these two opposite trends, the minimum
monthly temperatures (ATmin) do not show significant
changes in the mean over the period from January, 1981
through December, 2003. On the other hand, the maximum
monthly temperatures during the warmest days (MTmax)
increased noticeably from 10.7°C in the 1980s to almost 12.8°C
in the early 2000s (two unusual maximum daily temperatures
of 17.4 and 20.0°C were observed in May, 2001 and September,
2002, respectively). The maximum monthly temperatures
during the coldest days (mTmax) also increased from 5.1 to
6.1°C during the period from 1981 through 2003. As a
consequence of these two similar trends, the maximum
monthly temperatures (ATmax) exhibit a change from an
annual temperature of about 8.0°C in the 1980s to almost 9.4°C
in 2003.
The maximum monthly relative humidity records (MRH)
decreased from almost 100% in the 1980s to 96.5% in the early
2000s. The mean and minimum monthly relative humidity
values (ARH and mRH) also decreased from 96.1 to 88.2% and
from 83.8 to 70.6%, respectively, over the same period. In
several instances during the 1990s and the early 2000s, the
daily minimum relative humidity reached values below 60%
(July, 1994; January, 1997; June, 1998; February, 2000; and July,
2003). Finally, the maximum, mean and minimum diurnal
temperature range (MDTR, ADTR and mDTR) increased from
10.6 to 12.1°C, from 6.5 to 8.1°C, and from 3.1 to 4.2°C, res-
pectively, during the period from January, 1981 to December,
2003.
Besides the observed changes in the mean of the men-
tioned time series, the daily minimum sunshine values, the
maximum daily rainfall totals, and the maximum tempera-
tures during the warmest days exhibit increases in the
variance of their historical records. Moreover, the time series
of ATmin and mTmin show decreases in the variance during
the analyzed periods. Such changes are also evident in the
time series of the historical standard deviation (SD): max-
imum daily temperatures and daily relative humidity values
show evidence of increased SD, whereas minimum tempera-
tures exhibit decreased SD over the available 23-year periods.
3.2. Trends and potential driving mechanism
Based on the available period 1981–2003, records gathered at a
local weather station located on the west flank of El Ruiz–
Tolima volcanic massif at an altitude above 4000m, suggest
that minimum temperatures during the coldest days increased
at a rate of 0.6°C/decade (0.0055 ± 0.0018°C/month), whereas
maximum temperatures during the warmest and coldest days
increased at rates of about 1.3 and 0.6°C/decade (0.0109 ± 0.0017
and 0.0051 ± 0.0011°C/month), respectively. Estimated increas-
ing trends in temperature in low levels (nearby weather
stations at around 1300m) ranged between 0.2 and 0.4°C/
decade over the available period 1951–1970. The rates of in-
crease in maximum, mean and minimum diurnal temperature
range in the weather station located at an altitude above
4000m reached 0.6, 0.7, and 0.5°C/decade (0.0054 ± 0.0024,
0.0058 ± 0.0011 and 0.0040 ± 0.0011°C/month, respectively).
Maximum, mean and minimum relative humidity records
exhibit reductions of about 1.8, 3.9 and 6.6%/decade (− 0.0147 ±
0.0019, − 0.0328 ± 0.0036 and − 0.0550 ± 0.0080%/month,
respectively).
Increasing trends in temperature are likely to increase
precipitation in the form of rain particularly over the areas of
‘permanent’ snow, which used to have snowfall or only light
rain. Analyses suggest that, even though the local weather
station exhibits no changes in the mean of annual and
monthly rainfall totals during the period 1981–2003, it
evidences an increased occurrence of unusually heavy rainfall
events. One weather station located nearby at 3264m shows
a decreasing trend in annual rainfall of 4%/decade during
the period from 1950 through 1979. At altitudes around 1300m,
regional weather stations show annual rainfall totals
with decreasing trends of about 6%/decade over the period
1951–1970.
Changes in several other climatic variables have also occurred
in the selected high mountain region: observations suggest that
the reduction in the amount of all cloud types over the area
reached 1.9%/decade during the period from 1983 to 2001.
Accordingly, the total number of days per month of null sunshine
(foggy days) decreased significantly, and hence, the total number
of sunny days per month (TSD) increased by 2.1days over the
period from January, 1982 through December, 2005.
 S CIE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32
All these trends could be driven by changes in circulation
patterns in the Colombian Andes Mountains. As depicted in
Fig. 1, the Andes Cordillera is divided in Colombia into three
high branches (the Western, Central and Eastern Cordilleras)
by two large rivers, the Rio Cauca and the Rio Grande de La
Magdalena. The climatology of the Central Andean Colombian
region is strongly influenced by the behavior of three major
climate features: the Choco low-level westerly jet CLLJ (Poveda
and Mesa, 2000; Vernekar et al., 2003), which blows from the
Pacific Ocean and enters the country at almost 5°N and
between 850 and 1000mb; the Easterlies, which blow at al-
titudes above 4000–5000m; and the Northeastern winds,
which tend to blow from North to South along the valleys of
the Cauca and Magdalena rivers. The intensity of the CLLJ
depends on the difference between sea surface temperatures
in the Colombian Tropical Pacific Ocean (TPO) and El Nino 1 + 2
region (Poveda and Mesa, 2000). The inter-annual variability of
this low-level jet is dominated by the El Nino–Southern
Oscillation ENSO cycle: the CLLJs ‘is weaker in the warm
phase of ENSO than in the cold phase’ (Vernekar et al., 2003).
The dynamics and interactions between TPO, the CLLJ, the
Easterlies, the valley currents, and the orography of the Andes
Mountains at these latitudes lead to complex local scale
motions, considered the major mechanisms controlling the
formation of clouds and storms. The west upwind side of the
Andean Central mountain range (CMR), exhibits important
currents of lifting incoming air during the afternoon and
strong movements of masses of sinking air during nighttime.
The diurnal counter-clockwise cell formed over the west
flank of the Central Cordillera brings significant amounts of
water vapor from low to high levels in the atmosphere,
sometimes reaching altitudes of about 7000m. The moist
convection produced during this uplifting process plays a
significant role in the local climate effect of the high Andean
CMR. During the night, the upslope winds are reversed and
blow from the peaks to the Cauca valley lowlands. Such
clockwise cell encounters a counter-clockwise cell coming
from the opposite side of the valley, creating a zone of deep
Fig. 4 – High altitude water body in the headwaters of Santa Barbara and Las Juntas creeks, Claro River, ‘Los Nevados’ Natural
Park, Andean central mountain range, Colombia. High altitude water bodies exhibit a marked seasonality (picture on the left
was taken in September, 2005; picture on the right, two years later); the medium-to-long-term effect is, however, the
disappearance of the water body.
129
convection that is associated with the occurrence of heavy
storms over the Cauca River valley. These patterns are in
accordance with those by Vernekar et al. (2003), who analyzed
the variability of the lower tropospheric circulation and
precipitation on diurnal, intra-seasonal and inter-annual
timescales: they suggested that the CLLJ and the precipitation
patterns in the region show strong diurnal variability with
nocturnal maximums during the long summer period Janu-
ary–March. It is hypothesized, although not assessed here,
that the currents of lifting incoming air during the afternoons
are weakening due to changes in the intensity of the CLLJ and
the valley currents, and due to increases in atmospheric
temperatures and changes in vegetative cover of the West
flank. As a consequence, less water vapor is being produced
over cloud forests and less fog is reaching high altitudes on
these mountains. Hence, current scenario shows that fog
tends to be ‘trapped’ in low levels of the atmosphere, and only
high-level clouds (which have the net effect of increasing
surface temperatures) are ‘covering’ these high mountain
ecosystems. Changes in atmospheric stability and the lifting
condensation level are consequently expected in these areas.
4. Conclusions
Changes in the mean and the variance, and significant trends in
historical time series of minimum and maximum temperatures,
relative humidity, rainfall events and sunshine were observed
at the headwaters of the Claro River. Over the period from
January, 1981 through December, 2003 minimum and max-
imum temperatures during the coldest days increased at a rate
of about 0.6°C/decade, whereas maximum temperatures during
the warmest days increased at a rate of about 1.3°C/decade.
Interestingly, rates of increase in ambient temperatures in the
high mountain region were significantly higher than those
observed at lowland weather stations. Maximum, mean and
minimum diurnal temperature ranges also increased at rates of
130 SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32

about 0.6, 0.7, and 0.5°C/decade. Maximum, mean and mini-
mum relative humidity records, conversely, showed reductions
of about 1.8, 3.9 and 6.6%/decade. Finally, the headwaters of the
Claro River experienced no changes in the mean of rainfall
totals, contrary to what was observed in lowlands where,
generally speaking, it was detected a tendency towards drier
conditions. However, the high elevation area evidences an
increased occurrence of unusually heavy rainfall events during
the period 1981–2003.
Moreover, reductions in the amount of all cloud types of
about 1.9%/decade were observed over the area during the
period from 1983 to 2001. As a consequence of the reduction in
the number of foggy days, the total number of sunny days per
month increased by 2.1days over the period from January, 1982
through December, 2005. These decreases in the amount of
clouds could have contributed to the retreat of glacier icecaps, as
well as to changes in local albedo and radiation balance. The loss
of icecaps causes ‘abrupt changes in stream-flow, because of the
lack of glacial buffers during dry seasons’ (Bradley et al., 2006)
and because of the shrinking and disappearance of high altitude
water bodies and turberas (see pictures on Figs. 4 and 5). Changes
in the mean and the variance of minimum and maximum
temperatures, decreases in relative humidity, and increases in
the amount of days of significant incoming sunlight, lead to
slow upward shifts in the transition between life zones. On a
faster timescale, these climatic changes could favor the
occurrence and rapid spread of natural and man-induced forest
fires, which could undoubtedly disrupt the integrity of high
mountain ecosystems and cause dramatic losses of biodiversity
(see pictures on Fig. 6). In summary, all these striking alterations
and their intrinsic feedback mechanisms are threatening these
fragile ecosystems, causing sudden losses of unique mountain
species, and giving us steady signals of very adverse conditions.
High elevations are areas of primary concern because they
provide as much as 90–100% of the freshwater resources for
drinking, irrigation, and industrial supply in surrounding arid
and semi-arid lowlands, and because they are hotspots of
biodiversity (Diaz et al., 2003). Bradley et al. (2004 and 2006)
suggested that significant changes in high mountain regions
could be expected in the years to come, since temperatures in
these areas could rise more than those at lower elevations.
Should such predictions effectively occur a significant propor-
tion of Colombian communities are going to be seriously
affected: over 75% of Colombia's population lives in the
Andean region; 5million people out of these 34million persons
live in the surrounding areas of the Los Nevados Natural Park.

Fig. 5 – Permanent water-reservoir habitats (turberas) in the headwaters of Alfombrales Creek (top left), Claro River (top right and
bottom left), and Sietecuerales Creek (bottom right). The first three turberas still receive water supply from mountain ice caps
and high altitude water bodies. The Sietecuerales turbera used to receive supply from El Cisne ice-capped mountain, which
disappeared in the 1980s.
 S CIE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32
Fig. 6 – Los Frailejones (Espeletia hartwegiana) Valley, headwaters of the Claro River. Pictures were taken in March, 2006 (left) and
March, 2007 (right). These unique mountain species grow at a rate of few millimeters a year.
People are settled in several large cities, villages, and towns in
the immediate lowlands because those temperate areas
provide moderate climates that offer enjoyable living condi-
tions and because the rivers on the slopes of those mountains
support their drinking water demand. Moreover, the econo-
mies and activities of these communities depend directly on
these high mountain watersheds because they produce large
hydroelectric power potentials and they satisfy the water
demand of agricultural regions located in the intervening
Cauca and Magdalena valley lowlands. Accordingly, future
efforts will be focused on the economic assessment of the
environmental goods and services provided by the páramos of
the Los Nevados Natural Park, mainly water supply and basin
recharge, climate regulation, ecosystem integrity and biodi-
versity. Short and long-term economic impacts of the loss of
such natural resources due to the ongoing changing climate
should be quantitatively estimated through direct and indirect
assessments.
Undeniably, several gaps of knowledge remain in the
understanding of changes in atmospheric stability and their
role on the alteration of the local microclimate. A denser long-
term monitoring network is extremely needed to fulfill such
gaps and improve our ability to simulate and forecast future
climatic conditions. But we cannot wait until dramatic
reductions in water supply take place in the region. Policies,
strategies, and actions intended to ensure considerable levels
of environmental goods and services in Colombian high
mountain ecosystems need to be implemented immediately,
particularly because we are undoubtedly facing the ongoing
global changing climatic scenario. Fortunately, Colombia is
currently conducting an ambitious project aimed at defining
specific pilot adaptation measures and policy options to meet
the anticipated impacts of climate change on these areas.
Among its major goals, the project is conducting several
activities to maintain high levels of biodiversity in these
ecosystems, as well as to ensure considerable levels of water
supply, extremely needed to satisfy human consumption,
agriculture activities, industrial processes, and hydropower
generation. Moreover, the project is strengthening the cap-
ability of the Colombian government to produce and dis-
seminate continuous and reliable climate information
131
relevant to high mountain ecosystems. We are preparing
ourselves, but the scenario is definitely adverse.
Acknowledgements
We thank Adriana Maria Molina, Sandra Cristina Arias, and
Catalina Londoño Cadavid from Programa en Ingeniería
Ambiental—Escuela de Ingeniería de Antioquia, Arnold Gor-
don from the Lamont-Doherty Earth Observatory—Columbia
University in the City of New York, Alvaro Jaramillo from the
Centro Nacional de Investigaciones de Café—CENICAFE, and
Kelly Cunningham and Juan Esteban Quiroz Giraldo for all
their valuable comments. We thank the Unidad Adminis-
trativa Especial del Sistema de Parques Nacionales Naturales
(UAESPNN) de Colombia and the field team (Luis Fernando
Giraldo and Alirio Tibaguy) for all their support. We also wish
to thank two anonymous referees for their valuable sugges-
tions. Current activities are being supported by Dirección de
Investigación, Escuela de Ingeniería de Antioquia.
REFERENCES
Baker CB, Eischeid JK, Karl TR, Diaz HF. The quality control of
long-term climatological data using objective data analysis.
Preprints of AMS ninth conference on applied climatology,
January 15–20, Dallas, TX; 1995.
Bradley RS, Keimig FT, Diaz HF. Geophys Res Lett 2004;31:L16210.
Bradley RS, Vuille M, Diaz HF, Vergara W. Threats to water supplies
in the tropical Andes. Science 2006;312:1755–6.
Chaves B, Jaramillo A. Regionalization of air temperature in
Colombia. Revista Cenicafé 1998;24:91–104.
Díaz S, Fargione J, Stuart Chapin III F, Tilman D. Biodiversity loss
threatens human well-being. PLoS Biol 2006;4(8):e277.
doi:10.1371/journal.pbio.0040277
Diaz-Granados MA, Navarrete JD, Suarez T. Páramos:
hidrosistemas sensibles. Revista de Ingenieria, vol. 22. Bogota
(Colombia): Facultad de Ingenieria, Universidad de Los Andes;
2005. p. 64–75.
Diaz HF, Grosjean M, Graumlich L. Climate variability and change
in high elevation regions: past, present and future. In: Diaz HF,
132 SC IE N CE OF T H E TOT AL E N V I RO N ME N T 3 9 8 ( 2 00 8 ) 1 2 2–1 32
editor. Climate variability and change in high elevation
regions: past, present and future — advances in global change
research. Klumer Academic Publishers; 2003. 282 pp.
Euscategui C. Incidencias de las variaciones del brillo solar en la
dinámica glaciar del volcán Nevado de Santa Isabel. Revista
Meteorología Colombiana. Bogota: Universidad Nacional de
Colombia Sede; 2002.
Euscategui C, Ceballos J. El cambio climático sobre los glaciares
colombianos. V Congreso Colombiano de Meteorología, March
23–25, Bogota, Colombia; 1999.
Gutierrez ME, Zapata PA, Ruiz D. Understanding the signals of
climate change and climate variability on the water supply of
high mountain watersheds (in Spanish). Antioquia (Colombia):
Environmental Engineering Program, School of Engineering;
2006. 160 pp.
Kaser G, Georges Ch. Changes in the equilibrium line altitude in
the Tropical Cordillera Blanca (Peru) between 1930 and 1950
and their spatial variations. Ann Glaciol 1997;24:344–9.
Poveda G, Mesa OJ. On the existence of Lloró (the rainiest locality
on Earth): enhanced ocean–land–atmosphere interaction by a
low-level jet. Geophys Res Lett 2000;27(11):1675–8.
Poveda G, Mesa OJ, Velez JI. Balances hidrológicos de Colombia.
Universidad Nacional de Colombia Sede Medellín: Publicaciones
Postgrado en Aprovechamiento de Recursos Hidraulicos; 2001.
150 pp.
Rossow WB, Walker AW, Beuschel DE, Roiter MD. International
Satellite Cloud Climatology Project (ISCCP), Documentation of
new cloud datasets. World Meteorological Organization; 1996.
TD-No 737, 115 pp.
The World Bank Group. Colombia-Integrated National Adaptation
Program Project; 2006. Documents & Reports, World Bank
(http://www.worldbank.org/).
Vernekar AD, Kirtman BP, Fennessy MJ. Low-level jets and their
effects on the South American summer climate as simulated
by the NCEP Eta Model. J Climate 2003;16(2):297–311.
Vuille M, Bradley RS, Werner M, Keimig F. 20th century climate
change in the tropical Andes: observations and model results.
In: Diaz HF, editor. Climate variability and change in high
elevation regions: past, present and future — advances in global
change research. Klumer Academic Publishers; 2003. p. 282.
Wagnon P, Ribstein P, Francou B, Pouyaud B. Annual cycle of
energy balance of Zongo Glacier, Cordillera Real, Bolivia.
J Geophys Res 1999;104(D4):3907–24.
World Resources Institute. Millennium ecosystem
assessment—ecosystems and human well-being: biodiversity
synthesis; 2005. Island Press, Washington D.C., 86 pp. ISBN: 1-
59726-040-1.