USF Libraries
USF Digital Collections

Water quality in a tropical montane cloud forest watershed, Monteverde, Costa Rica


Material Information

Water quality in a tropical montane cloud forest watershed, Monteverde, Costa Rica
Translated Title:
Calidad de agua en una cuenca en el bosque nuboso tropical, Monteverde, Costa Rica ( )
Physical Description:
Rhodes, Amy L
Place of Publication:


Subjects / Keywords:
Water quality
Calidad de agua
Books / Reports / Directories   ( local )
Books / Reports / Directories   ( local )



Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - M36-00398-ML-1003
usfldc handle - m36.398-ml-1003
System ID:

This item is only available as the following downloads:

Full Text


Rhodes et al. page 1 of 12 WATER QUALITY IN A TROPICAL MO NTANE CLOUD FOREST WATERSHED, MONTEVERDE, COSTA RICA Amy L. Rhodes1,3, Andrew J. Guswa2,3, Stewart Dallas3, Evelyn M. Kim1, Sarah Katchpole1, and Ann Pufall1 (1) Department of Geology, Smith College, Northampton, MA, USA 01063, (2) Picker Engineering Program, Sm ith College, Northampton, MA, USA 01063 (3) Monteverde Institute, Monteverde, Puntarenas, Costa Rica Citation: Rhodes, A.L., A.J. Guswa, Dallas, S., Kim, E.M ., Katchpole, S., Pufall, A., in press. Water quality in a tropical montane cloud forest watershed, Monteverde, Costa Rica. In: Bruijnzeel, L.A., Juvik, J., Scatena, F. N., Hamilton, L.S., and Bubb. P. (eds), Mountains in the Mist: Science for Conservatio n and Management of Tropical Montane Cloud Forests, University of Hawaii Press.


Rhodes et al. page 2 of 12 Abstract Monteverde, Costa Rica is home to a tropical mont ane cloud forest that illustrates the balances among habitat and development inhe rent to ecotourism. The Mont everde Cloud Forest Preserve, located on the leeward side of the continen tal divide, has experienced a one-hundred-fold increase in visitors since its in ception twenty years ago. The associated growth in population and commercial development has the potential to imp act water resources. Over three years, we collected more than four-hundred stream samples from eight s ites above and below the main road within the Rio Guacimal watershed to asse ss the effect of developm ent on water quality. The chemistry of upstream samples is characteri stic of mineral weathering and cation exchange reactions in soils. Comparisons of downstream samples to these baseline data show evidence of anthropogenic impacts; chemical loads are two to five times more concentrated at the downstream locations. Highest co ncentrations are seen at QS-200, the site with the highest population density. These results point to the value of forest preserves, specifically the Monteverde Cloud Forest Reserve and the Bajo de Tigre, in limiting growth in riparian regions, which in turn helps to protect wate r resources for downstream communities. Introduction Tropical montane cloud forests (TMCFs) are ecosy stems of extraordinary biological diversity whose existence depends on frequent immersion in clouds and mist. M onteverde, Costa Rica harbors highland TMCFs that exemplify the delic ate balances among climat e, hydrology, habitat, and development. This region has received signifi cant attention due to its high biodiversity (cf. Nadkarni and Wheelwright, 2000), and it is undergoing rapid de velopment as an ecotourist destination. The success of ecotourism has spurre d the construction of new hotels, restaurants, businesses, and homes in Monteverde and th e neighboring communities of Cerro Plano and Santa Elena. Water plays a prominent role in Monteverde; it is essential for the local ecosystem as well as the community. Drinking water is taken from springs and streams in the cloud forest reserve, and increased demands for water will re duce streamflow. This could affect aquatic habitats and limit the ability of these streams to dilute graywater discharge and agricultural runoff. Decreased cloud formation in the regi on brought on by climate change (Still et al., 1999; Lawton et al., 2001; Foster et al., this issue) c ould further stress the quality and quantity of the


Rhodes et al. page 3 of 12 water. This study characterizes the water chemistry of a 19.7 km2, headwater catchment of the Rio Guacimal to illustrate the impacts of land use on water quality. Study Area Monteverde (108N, 8448W; 1500 m) is located on the leeward (western and Pacific) side of the Cordillera de Tilarn, a stre tch of the volcanic range that defines the continental divide of Costa Rica (Figure 1). The climate of this region is marked by three distinct seasons, which are related to the north-south migr ation of the Intertropical Conve rgence Zone (ITCZ) across Costa Rica (Clark et al., 1998; Clark et al., 2000). Annually, Monteverde receives ~2519 mm of precipitation each year (Clark et al., 2000), with ~70% of annual precipitation occurring during the wet season (May-October) when the ITCZ is positioned directly over Monteverde. The transitional (November-January ) and dry (February-April) seasons correspond to the months when the ITCZ is located to the south of Co sta Rica. Average precipitation in Monteverde during the dry season is less than 50 mm per month. In comparis on, the average annual rainfall at sea level further leeward, as measured at Palo Verde Biological Station (Organization for Tropical Studies, 2005), is 2.5 tim es less than the annual rainfall at Monteverde, suggesting that the headwater catchments of the Monteverde re gion are important suppliers of water to downstream communities on the Pacific slope. Much work is currently ongoing to better understand the hydrology of the Monteverde region (e .g., Dalitz et al., this i ssue; Frumau et al., this issue; Kohler et al ., this issue; Rhodes et al., this issu e; Schmid et al., this issue; TobnMarin et al., this issue). This study focuses on the headwater catchment (1970 ha) of the Rio Guacimal (Figure 1), which originates in the Monteverde Cloud Fo rest Reserve and extends from 1842 m on the continental divide to its conf luence with the San Luis River (elevation 700 m). The upper elevations of the watershed are forested and la rgely undisturbed by human activity; they include primary and secondary forest as well as forest preserves (Monteverde Cloud Forest Reserve and Bajo de Tigre). Development is concentrated in the towns of Santa Elena and Cerro Plano, and continues along the main road that leads to the Monteverde Cloud Fo rest Reserve. Agricultural


Rhodes et al. page 4 of 12 activity—farms for dairy cows, pi gs, coffee, and other crops—is concentrated below the main road. Approach Since 2000, water samples have been collected th roughout the watershed and analyzed for major ion chemistry. This work presen ts the results of 432 water samples collected from eight sites, which represent remote and populated areas (Figure 1). Three sites (QC-300, QM-300, and HB100) are located on first-order streams that drai n forested, upland regions, and they serve as reference locations for assessing human influen ces on water quality further downstream. Three sites (RG-200, QM-200, QS-200) are located just downstream of the main road to the cloud forest reserve that bisects the watershed. Of these, QS-200 drains the most commercial and densely populated area; all three have similar climate and topography. At lower elevations, the watershed becomes warmer and drier. QSO-100 (~940 m) receives runoff from pasture and lowdensity residential areas The lowermost (700 m) sampling point (RG-100) receives water from all these sites and runoff from larger-scale agricu lture, including a commercial pig farm. Stream water samples were collected, on average, every 24 days, although the number of days between collections varied between 2 and 74 days. Rain fall and throughfall sample s also were collected in a clearing and under the canopy of sec ondary forest during July – October 2003. All water samples were collected in acid-washed polyethylene bottles, after rinsing three times with the sample, and then refrigerated. Samples were analyzed at the Aqueous Geochemistry laboratory at Smith College fo r pH, specific conductance, acid neutralizing capacity (ANC), cation concentrations (Ca2+, Mg2+, Na+, K+, NH4 +), anion concentrations (F-, Cl-, NO3 -, PO4 3-, SO4 2-), and dissolved silica. ANC was measured by Gran titration on unfiltered samples. Samples then were vacuum filte red through a 0.45 micron Millipore membrane. Cation concentrations of acidifi ed splits and anion concentrati ons of unacidified splits were determined by ion chromatogra phy, using a Dionex 500 ion chroma tograph. Instruments were standardized to solutions of known concentrations ; errors in accuracy are <0.13 mg/L for cations and <0.18 mg/L for anions. Dissolved silica c oncentrations were determined by colorimetry using a Bauch and Lomb Spectronic 21 spectrometer. For each sample, charge balances were calculated in equivalents as ( cations – anions)/( cations + anions)*100%. Samples having imbalances greater than 5% were either reanalyzed or assessed for cause of the imbalance. We


Rhodes et al. page 5 of 12 estimate an average ANC error of 20 eq/L from the charge balanc e calculations. Results are reported in molar (mol/L) units; equivalents (eq /L) are used when data interpretation requires charge balances. Due to the sc ope of this paper, dissolved si lica results are not presented. Impacts of land use on stream water chem istry are evaluated by comparing samples collected from the upland, forested streams to those suspected to be impacted by development. In an undisturbed region, we expect the dominan t controls on stream water chemistry to be mineral weathering and cation exchange reacti ons between precipitation and soil and rock materials. Base cations (Ca2+, Mg2+, Na+, K+) and alkalinity (measured as ANC) are added in equal amounts to streamwater by these processes. Pointa nd nonpoint-source pollution may add strong acids (indicated by SO4 2and NO3 -) and salts (identified by Cl-). Graywater discharges and runoff containing fertilizers may also cont ribute alkalinity, which increases the ANC and base-cation concentrations of th e streams (Rhodes et al., 2001). To evaluate whether anthropogenic compounds fr om the atmosphere are affecting stream chemistry, we compare the chemistry of throughfa ll and open precipitation to the chemistry of streamwater in the cloud forest. The impact of pollution added by local land-use activities is determined by comparing sites below the main ro ad against the remote forested sites. We investigate possible correlations between the sum of base cations (SBC = Ca2++ Mg2++ Na++ K+) and the sum of acid and salt anions (SAA = SO4 2+ NO3 + Cl-) for each sample site. A weak correlation between SBC and SAA would be cons istent with natural processes having the strongest control on stream water chemistry; weathering of minerals and soils adds far more base cations than acid anions. We expect that polluti on from multiple land uses will show a stronger correlation between SBC and SAA. Also, we expect the pollution signatures to exhibit greater variability as the intensity of human activity is not constant th roughout the year. The seasonality of precipitation also plays a st rong role in either adding (thr ough runoff) or diluting pollution. Mineral weathering by carbonic ac id and cation exchange reac tions in soils leads to a one-to-one increase in SBC and ANC. SBC exceeds ANC when strong acids reduce alkalinity or salts add base cations. In the forested sites, we expect SBC to be sli ghtly higher than ANC due to acid rain and naturally occurring organic acid s from the environment. Larger differences between SBC and ANC than are observed at the fore sted sites quantify the impact of pollution on the other streams.


Rhodes et al. page 6 of 12 Results and Discussion The chemical composition of open precipitation (Figur e 2) is consistent with the range of results reported for Monteverde cloud forest s by Clark et al. (1998). Aver age concentrations of sodium and chloride (17 mol/L each) suggest that marine ae rosols have a strong influence on precipitation chemistry. Average sulf ate and nitrate concentrations (7.1 mol/L and 2.2 mol/L, respectively) are lower than the northeastern United States (NADP 2002 data, accessed 2004), where rainfall is affected by indus trial air pollution. Sulfate in rain that reaches Monteverde could be derived from volcanic emissions; Vo lcan Arenal is located ~25 km upwind. Interestingly, the precipitation collected is not acidic (pH = 5.5 – 6.1; ANC = 15 – 47 eq/L) and the average concentration of th e non-salt base cations (Ca2+ + Mg2+ + K+ = 21.7 mol/L) is higher than what is seen in the majority of precipitation across the U.S. (NADP 2002 data, accessed 2004). In Monteverde, Clark et al. (1998) observed much higher concentrations of all ions in cloudwater and mist than in rain, and th ey attributed cation and n itrate concentrations to burning of agricultural and forest biomass in lowl and areas of Costa Rica. We suspect that the atmospheric mixing of precipitation with dus t and ash — whether fr om biomass-burning, volcanoes, or dirt roads — neutralizes the acidity normally observed in rain. Differences in chemistry between open precipitation and throughfall likely are due to a combination of effects. Wash off of atmospheric material accumulate d on plant surfaces by dry deposition and direct evaporation of intercepted water would increase ion concentrati ons. Exchange processes that occur between precipitation and vegetation could add K+, Ca2+ and NH4 + to throughfall (Hansen et al., 1994). As indicated in Figure 2, aver age concentrations of major ions in streamwater increase with decreasing elevation and increasing land use. The upland streams are 1.5-3 times more concentrated than open precipita tion and throughfall due to a comb ination of mineral weathering and evapotranspiration. The chemical loads of streams receiving runoff from populated areas are 2-5 times more concentrated than the upland st reamwater sites. Among the upland sites (HB100, QM-300, and QC-300), Na+ > Ca2+ > Mg2+ > K+ >> NH4 +. The ratios between cations are the same at these sites, suggesting that bedroc k mineralogy and soil chemistry in the Q. Maquina


Rhodes et al. page 7 of 12 and Q. Cuecha catchments are similar. The cati on results are consistent with the andesitic, volcanic bedrock composition of the area and th e base saturation (48-99%) of organic and mineral horizons in soils being high (Kim et al., 2002). QM-300 and HB-100 have higher overall base cation and ANC con centrations than QC-300, suggesti ng that subsurface water in the Q. Maquina catchment has a deeper flowpath and a longer contact time with soil and rock materials before entering the stream channel. For acid and salt anions, chloride concentration is greatest (Cl> SO4 2NO3 >> PO4 3-) and varies little among the three upland sites. In combination with the observed variation in cations, this suggests that the minerals ar e not a significant sour ce of chloride. SO4:Cl ratios in these streams are higher th an in precipitation, and, lik e base cations and ANC, SO4 is greater at QM-300 and HB-100 than QC-300. This indicat es that mineral w eathering — likely of accessory sulfide minerals commonly found in volcanic rocks — adds sulfate to the streams. Strong acid and salt anion concen trations of the Q. Maquina and Q. Cuecha increase only slightly downstream of the main road. Elevated concentrations of chloride and sodium at RG200 (Q. Cuecha is renamed Rio Guacimal at the road) are observed occasionally. We attribute this rise to salts in cheese-processing waste di scharged by the Monteverde Cheese Factory, located 100 meters upstream of our sample site. Ov erall, the forest preser ve, Bajo de Tigre, and other adjacent forests has limited development along the riparian zone of the upper reaches of these streams, which minimizes nonpoint source pollution. In stark contrast, the Q. Su cia (site QS-200), which drains the most densely populated area in the watershed, shows much higher concentrations of all i ons. Plotting the sum of base cations (SBC) against the sum of acid and sa lt anions (SAA) shows that pollution strongly controls the water chemis try of Q. Sucia (Figure 3). Acid anion concentra tions (SAA = 150-400 eq/L) are more than twice as high as what is observed at the upland sites (SAA = 75-150 eq/L). The wide range in SBC (550-1350 eq/L) combined with higher average ANC (616 eq/L) indicates that alkaline pollution is added to Q. Sucia as well. In contrast to QS-200, the elevated concen trations observed in Q. Socorro cannot be attributed to high inputs of pollu tion. The range of SAA (125-200 eq/L) is slightly elevated but


Rhodes et al. page 8 of 12 comparable to what is seen in the upland sites (SAA = 75-150 eq/L). Also, while SBC is high (900-1150 eq/L), its variability is comparable to th e samples from the upland sites. These observations suggest that mineral weathering a nd cation exchange are the major controlling processes on the chemistry of Q. Socorro, and de velopment has a minimal impact on this stream. Changes in bedrock composition, warmer temperat ures, and longer residen ce time of subsurface flow at this lower elevation site could a ll explain the higher conc entrations, although the influence of fertilizers on ba se cation concentrations and alkalinity cannot be ruled out completely. The net effect of adding acids, salts, and alkalinity to stream s is shown in Figure 4. If base mineral weathering and cation exchange reactions were the only controls on stream water chemistry, a theoretical 1:1 relationship betw een SBC and ANC, havi ng an intercept = 0 eq/L, would exist. While alkaline pollu tion does not change the observ ed 1:1 relationship, acids and salts (which remove ANC and add base cations, respectively) shift values away from this theoretical line (Rhodes et al., 2001). On av erage, atmospheric effects on precipitation and sulfide weathering account for approximately an 80 eq/L shift in the intercept of SBC versus ANC at the remote forested sites. Relative to this baseline, further deviations from the theoretical relationship between SBC and ANC quantify the impact of pollution. Figure 4 shows clearly that the Q. Sucia and th e lower Rio Guacimal are the most impacted. In all eight sites, acidic pollution (SO4 2+ NO3 -) is approximately equal to salt additions (Cl-), and these concentrations are greates t at QS-200 and RG-100. Conclusions Comparison of stream chemistry upstream and downstream of development in Monteverde, Costa Rica shows evidence of impacts to wate r quality arising from urban and agricultural development. Among the sites located just do wnstream of the main road, the anthropogenic signature is strongest at QS-200. The chemistry at RG-200 and QM -200 is quite similar to that of the upland sites. These results point to the va lue of the preserves, specifically the Monteverde Cloud Forest Reserve and Bajo de Tigre, in limiting development and associated point and nonpont source pollution to recharge regions of watersheds. Future work will seek to elucidate


Rhodes et al. page 9 of 12 details of mineral weathering, causes behind tempor al variations in stream chemistry, and links between hydrologic variation a nd geochemical variability. Acknowledgements We thank Nat Scrimshaw, The Monteverde Inst itute, and the Monteverde Community Arts Center for providing logistical support and staff resources. B. Scheffe, A. Harmon, M. Marin, J. Torres, and J. Neibler, all of MV I, assisted with stream sampling. S. Newell, C. Chazen, and M. Mick assisted with chemical analysis. Funding was provided by Smith College faculty development grants and the Smith Su mmer Science Internship Program. References Cited Clark, K., R. Lawton and P. Butler (2000). Chapter 2: The Physical Envi ronment. Monteverde: Ecology of and Conservation of a Tropical Cloud Forest N. M. Nadkarni and N. T. Wheelwright, (Eds.) Oxford University Press, New York, p. 573. Clark, K., N. Nadkarni, D. Schaefer and H. Gholz (1998). Cloud wa ter and precipitation chemistry in a tropical montane forest, Mont everde, Costa Rica. Atmospheric Environment 32 (9): 1595-1603. Dalitz, H., M. Oesker, H. Todt, and A. Wolter, th is issue. Spatio-temporal heterogeneity of canopy throughfall in montane and upland rain forests of Costa Rica, Ecuador and Kenya – A comparison of variation and biodiversity. Foster, P., S. Schneider, J. Brad ford, and M. Silman, this issue. Climate change induced changes in the altitudinal range of cloud formation. Frumau, A., S. Bruijnzeel, C. Tobn-Marin, and J. Calvo, this issue. Cl oud forest and pasture water use in northern Costa Rica.


Rhodes et al. page 10 of 12 Hansen, K., G. Draaijers, W. Ivens, P. Gunde rsen, and N. van Leeuwen, 1994). Concentration variations in rain a nd canopy throughfall collec ted sequentially during in dividual rain events. Atmospheric Environment 28 (20): 3195-3205. Kim, E., Rhodes, A., Katchpole, S., Wells, A., Scheffe, B., Dallas, S., Pufall, A., 2002. Water quality study of a cloudforest watershed in Monteverde, Costa Rica. The Geological Society of America, 2002 Northeastern Section Annual Meeting, Abstra cts with Programs v. 34, p. 18. Kohler, L., D. Hlscher, C. Tobn-Marin, and S. Br uijnzeel, this issue. Epiphyte distribution and water dynamics in old-growth and secondary montane forests under contrasting climatic conditions in Costa Rica. Lawton, R., U. Nair, R. Pielke and R. Welc h, 2001. Climatic impact of tropical lowland deforestation on nearby montan e cloud forests. Science 294 : 584-587. Nadkarni, N. and N. Wheelwright, 2000. Chapter 1: Introduction. Monteverde: Ecology and Conservation of a Tropical Cloud Forest N. Nadkarni and N. Wheel wright, (Eds.) Oxford University Press, New York, p. 573. National Atmospheric Deposition Program (NADP ) (accessed 5/9/2004) 2002 Annual Isopleth Map, Organization for Tropical Studies (accessed 2005) Palo Verde Metereological Data, Rhodes, A., R. Newton, A. Pufall, 2001. Influen ces of land use on water quality of a diverse New England watershed. Environmental Science and Technology 35 : 3640-3645. Rhodes, A., A. Guswa, S. Newell, this issue. Using stable isotopes to identify orographic precipitation events in Monteverde, Costa Rica.


Rhodes et al. page 11 of 12 Schmid, S., R. Bukard, A. Fumau, C. Tobn-Marin, S. Bruijnzeel, R. Sieg wolf, and W. Eugster, this issue. The wet-canopy water ba lance of a Costa Rican cloud forest. Still, C. P. Foster and S. Schneider, 1999. Simu lating the effects of climate change on tropical montane cloud forests. Nature 389 : 608-610. Tobn-Marin, C., S. Bruijnzeel, A. Frumau, and J. Calvo, this issue. Changes in soil hydraulic properties and soil water status af ter conversion of tropical montan e cloud forest to pasture in northern Costa Rica.


Rhodes et al. page 12 of 12 Figure Captions: Figure 1: The Rio Guacimal watershed and sample locations. Figure 2: Average (A) cation and (B) anion concen trations of precipitation and stream samples. Open = open precitation; TF = throughfall. Figure 3: Sum of acid and salt anions (SAA) vs. sum of base cations (SBC). Figure 4: SBC vs. ANC quantifies the impact of pollution at each sampling location. Acidic pollution (SO4 2+ NO3 -) is approximately equal to salt (Cl-) additions.


500 m E HB-100 QS-200 QM-300 RG-200 QC-300 QSO-100 RG-100 QM-200 Q. Cuecha Rio Guacim al Q. Cambronero Q. Socorro Q. Maquina Q. Sucia Monteverde Cerro Plano Santa Elena Lege nd Roads Watersh ed Boundary Riv e rs San Jose Monteverde P ACIFIC OCEAN CARIBBEAN SEA 0 60 km N N Precipitation Site Cloud Forest Preserve En tra n ce E R. Guacimal


0 100 200 300 400 500 0 200 400 600 800 1000 1200 1400 Upland Streams QSO-100 QS-200 Open TF SAA (meq/L) SBC (meq/L) R2 = 0.35 R2 = 0.26 R2 = 0.74


0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 SBC, Upland SBC, Below Road SO4 + NO3 Cl y = 81 + 1.01x R 2 = 0.997 SBC, SO4 + NO3, Cl ( eq/L) ANC ( eq/L) RG-200 QM-300 QS-200 Theoretical 1:1, SBC vs. ANC RG-100 QSO-100

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader 00000nas 2200000Ka 4500
controlfield tag 008 s2005 xx |||| |||||| eng
datafield ind1 7 ind2 024
subfield code a M36-00398-ML-1003
1 100
Rhodes, Amy L.
0 242
Calidad de agua en una cuenca en el bosque nuboso tropical, Monteverde, Costa Rica.
Water quality in a tropical montane cloud forest watershed, Monteverde, Costa Rica.
g 2005?
Books / Reports / Directories
2 local
Water quality
Calidad de agua
Scanned by Monteverde Institute.
Amy L. Rhodes and Andrew J. Guswa
The State of Water in Monteverde, Costa Rica: A Resource Inventory.
4 856