Supervisor : Mark Mulligan

INTRODUCTION

The present rate of deforestation in Colombia is 600,000 ha/year (D.N.P, 1994). After being deforested, soils are used for agriculture or grazing, although in most cases these lands do not have vocation for these activities. This is causing serious environmental and social problems. Biodiversity reduction, erosion, sedimentation of water bodies and lack of regulation of water yields have brought serious ecological and economical costs.

One of the most heavily affected ecosystems in the Andean region is the tropical mountain cloud forest (TMCF). Only 10% of the original TMCF of the Andes is left, largely because approximately 56% of the Colombian population originally settled in the Andean mountains between 1,500 and 3,000 m (Castaño, 1991).

TMCF has been recognised as particularly important in supplying water during rainless periods (Zadroga, 1981, Hamilton and King, 1983, Brujinzeel and Proctol, 1995). Traditionally accepted inter-relationships between land-use and watershed hydrology do not apply to TMCF ecosystems. It has been suggested that the removal of TMCF could reduce the water yields due to decreased net precipitation. In these forests, precipitation may occur as moisture intercepted by the vegetation, or occult precipitation, as well as by rainfall (Bruijnzeel and Proctor, 1995).

Colombian water utilities and environmental agencies have identified the need to preserve water resources and are developing a series of economic incentives, such as subsidies and tax reductions, to preserve the cloud forest. These types of incentives could benefit small landowners and farmers who could contribute to forest conservation while receiving an economical benefit for the supply of non-wood forest products, such potable, well-regulated water. However, the implementation of these incentives has been limited by the lack of knowledge of the factors controlling the relationships between different TMCFs, land use changes and watershed dynamics (DNP, 1994).

Lawrence et al. (1995) defined TMCF as an ecosystem with distinctive floristic composition where persistent or seasonal cloud cover occurs at the vegetation level. Cloud forest occurs in a broad range of altitudes depending on the size of the mountain, its distance from the ocean and exposure to prevailing winds. In some areas, the cloud belt is located at 1,500 m and in others it can be found at 3,900 m.

TMCF water budget is highly affected by the persistent cloud cover. The fog results not only in additional water input, but also in an environment with reduced solar radiation, lower vapour pressure deficit and higher humidity. These atmospheric characteristics normally lead to low evapotranspiration rates and wet canopies (Lawrence et al., 1995). In some areas, the contribution of cloud intercepted by the vegetation to the total water input of the catchment could be 50%. This is the case of the TMCF of Parque Nacional Natural Macuira in northern Guajira, Colombia (Cavelier, 1991). However, in other areas, such as Zumbador, Venezuela, the contribution of occult precipitation to the water input as low as 3.5%. In the case of Macuira, not only is there less rainfall, but there is generally a higher water content in the ground-level cloud than in Zumbador. Some researchers have suggested that water-laden cumuliform clouds are more common in Macuira than in the higher northern Andes, where finer stratiform clouds are prevalent (Cavelier and Goldstein, 1989).

Despite the importance of occult precipitation in the water budget of TMCF its quantification is difficult, not only for the lack of standardised automated instruments to measure horizontal precipitation, but also for the lack of knowledge of the processes controlling cloud interception in natural environments. Kerfoot (1968) was the first to review the importance of cloud interception in the hydrology of TMCF and also the first to investigate the problems of measuring horizontal precipitation. He identified two approaches to quantify horizontal precipitation. The first is the use of cloud interceptors, which gives a good indication of relative moisture contributions of low clouds along a catchment. The problem with these devices is that there are no known relationships between these interceptors and the forest vegetation, especially given that each forest community is unique. The second approach is the comparison of throughfall inside and outside of a stand. These types of measurements provide a site-specific estimation of cloud interception, if the interception, stemflow and evapotranspiration are also quantified.

Bruijnzeel and Proctor (1995), in a recent, comprehensive review, illustrated that some vegetation and climatic factors could control the processes of cloud interception. Vegetation characteristics that can affect cloud interception are height, canopy structure, canopy size, biomass, orientation and physical characteristics of leaves and epiphytes. Climatic factors include moisture content, drop size, wind velocity and direction, slope, topographic position and duration of cloud cover. However, horizontal precipitation varies greatly between different TMCF and the role of vegetation in the dynamics of the TMCF environment and in the process of cloud interception is still unclear (Bruijnzeel and Proctor 1994, Cavelier et al 1996).

This study will serve to improve our understanding of the interrelationships between land use changes, climate and hydrological responses in small catchments in the southern Colombian TMCF.

The aim of this study is to further understand the processes of occult precipitation, its contribution to the water balance of Andean catchments and possible implications of forest removal.

• To explore the role of southern Colombian TMCF vegetation in occult precipitation, through an estimation of the interception efficiency of dominant and abundant families of trees and epiphytes and to identify possible implications of forest removal.
• To determine the atmospheric and biotic characteristics controlling the process of cloud moisture interception in the TMCF of southwestern Andean Colombia and to determine their frequency and variability along altitudinal gradients and along the wind profile.
• To estimate the contribution of occult precipitation to the water budget of two small cloud forest catchments in southern Colombia and determine its importance during rainless periods.
• To test (compare/contrast) various instruments for the measurement of occult precipitation with a view to quantify their suitability and to correlate their cloud interception efficiency against that of the native vegetation.

The study site is located in the Centro de Estudios Ambientales del Pacífico Tambito. This private forest reserve is located at 76º25’ W, 2º55´ N, approximately, 60 kilometres from the Pacífic Coast and 150 kilometers from El Tambo, Cauca. Following the Holdridge classification, the vegetation of the area is mainly Low Mountain Forest (LMF). The study will be conducted in the Tambito and Palo Verde catchments, ranging from 1,400 m where the two catchments meet, to approximately 2,200 m in the uppermost regions. The average annual temperature of the area varies from 13ºC to19ºC. Annual average rainfall ranges from 3800 to 7000mm. The catchments have areas of approximately 1,500 ha each and have similar soils and geology. The most significant difference between these two catchments is in their land-use history. While the Palo Verde catchment is composed of approximately 90% primary forest, 18% secondary forest and 2% pasture, Tambito has 10% primary forest, 70% secondary forest and 10% pasture.

• The dominant families of the primary and secondary forest will be classified according to functional types. Functional type will be defined by texture, surface area and architectural characteristics, which determine their cloud interception efficiency. Taxonomic analysis will investigate the distribution of the families and genera belonging to each functional type.
• Samples of each functional type will be tested in the field and laboratory to determine their water retention and cloud interception capacity. In the field, samples of different functional types will be exposed to the fog for a determinate time and changes in their weight will be recorded. In a closed chamber, individuals of each functional type will be exposed to different amounts of fog. The fog will be passed through the chamber until the sample is saturated.
• Regression relationships between retention capacity and leaf area will be determined. The leaf area index (LAI) will be used to scale the retention capacity of each functional type to the forest stand scale.

• At the catchment scale the dominant families will be identified and their abundance, density and basal area will be estimated using random point methods along an altitudinal gradient. Functional type, interception efficiency, biomass and LAI will be used to scale the potential interception efficiency of the vegetation along the gradient.
• Samples of epiphytes will be collected along altitudinal gradients in the catchments and their water retention capacity will be estimated in the laboratory and in the field. Estimates of the epiphytic biomass of small plots will be used to scale the contribution of epiphytes to net water storage.

• Over the course of a full year, the water balances of three permanent plots with different land uses will be monitored. Rainfall, throughflow, stemflow, soil moisture, overland flow, and subsurface flow records will be used to examine the water balances during wet and dry periods. Evapotranspiration will be calculated using the Penman-Monteith equation for wet and dry canopies. Cloud interception will be determined as a residual of the water balance. These values will be compared with the cloud water intercepted with the cloud catchers.

• Direct measurements of cloud-water content will be taken. Different cloud interceptors and humidity sensors will be studied to determine the best way to calculate the input of occult precipitation to the water budget. Examples of such instruments include gauze cylinders (Nagel, 1964), wire harps, and scales to measure the changing weights of absorptive materials such as sponges, vegetation and silica gel. Wind speed, temperature and solar radiation measurements will be evaluated as indicators of cloud moisture content, cloud distribution and frequency.
• Cloud interceptors will be placed at different locations within the catchment to account for altitudinal, spatial and topographic variations in cloud moisture content and at different heights within the canopy, to account for differences along the wind profile.

• For one year, streamflow, rainfall and occult precipitation will be taken in the two catchments in order to estimate water balance
• During rain-free periods, hydrograph data will be analysed to determine the significance of occult precipitation in the water budget

• After filtering long-term cycles (e.g. previous rainfall events), low flow daily cycles will be analysed to identify cloud interception signals that could be used to estimate the contribution of cloud interception to the local water budget
• The paired catchments will be compared to explore the role of vegetation in the hourly hydrograph signals
• A model will be designed and tested to estimate cloud interception in the Palo Verde and Tambito catchments