Climatic elements and factors

Horizontal precipitation and its measurement. It has already been necessary on various occasions to point out the influence of climate and its parameters on the presence, distribution and structure of cloud forests. Holdridge (1971) summarizes the climatic conditions necessary for cloud forests, naming them "special atmospheric conditions":

Wherever fogs and mists occur with great frequency, as they do on windward mountain slopes in the condensation or "cloud belt"... they may constitute a significant source of additional moisture. Fog-borne moisture, dew, and heavy mists may condense upon exposed vegetational surfaces, and drip, or run down stems, to the ground. Such moisture, however, is not recorded in properly installed standard rain gauges, and its quantity is known to be highly dependent upon both the successional stage and foliage characteristics of the dominant vegetation. Hence, it does not enter into the computation of mean and annual precipitation for the determination of the Life Zone itself. Rather, it is considered to be a wet-atmospheric factor entering the classification at the secondary or association level.

Many authors consider that the additional supply of water, in combination with lower temperatures in mountainous zones, is an important ecological factor affecting the different types of cloud forests. This input of precipitation, in addition to rain, has been variously named by different authors (Kittredge, 1948; Geiger, 1961; Lamb, 1965; Kerfoot, 1968; Holdridge, 1971; Whitmore, 1975; Caceres, 1981):

• horizontal interception
• negative interception
• cloud moisture interception
• fog precipitation ("Nebelniederschlag")
• fog drip/cloud drip
• condensation drip
• mist precipitation
• fog stripping
• occult precipitation
• horizontal precipitation

In the present study, the entry of water into the ecosystem conditioned by the condensation process of the humidity of clouds or fog on vegetational surfaces, or by means of direct contact of cloud droplets with the vegetation, will be called horizontal precipitation and signifies an input of additional water to rainfall. The term "horizontal interception" used in different publications could be misleading in that "interception" as a micrometeorological and hydrological parameter, is generally understood to mean water intercepted by vegetation and then evaporated from the plant surfaces; in other words, water leaving the ecosystem.

The quantity of horizontal precipitation depends as much on inherent vegetational factors as climatic factors and elements. These will be discussed later, when the different climatological scales (macro, meso and micro) will be analysed.

The inherent vegetational characteristics, are roughly the following:

• height of the vegetation
• canopy structure (influencing the roughness thus causing micro turbulence)
• size, quantity, location and arrangement of leaves
• quantity, forms and types of epiphytes.
• According to Juvik and Ekern (1978) two basic methods exist for the measurement of horizontal precipitation:

a) collect total precipitation beneath the canopy by means of troughs or a large number of rain gauges, subtracting from this the rainfall above the forest or in a nearby unforested area.

b) collect cloud droplets by means of artificial apparatuses known as "fog catchers".

The first method gives real values for net precipitation, being ecologically and hydrologically relevant for the ecosystem under study. Considering the interception (water intercepted by vegetation and then evaporated) the behaviour of gross rainfall (measured outside or above the forest) together with other climatic elements, this method allows the calculation of the actual contribution of horizontal precipitation to net precipitation. Measurement techniques require a large quantity of equipment, and are thus costly and present a difficulty in that the location of the collection troughs or rain gauges does not give readings that are representative of the ecosystem as a whole.

The second method that has been used by many scientists outside the tropics (see chapter 1) and by Ekern (1964), Baynton (1969), Vogelmann (1973), Juvik and Ekern (1978) and Caceres (1981) in humid tropical zones, determines the quantity of water that can theoretically be extracted from clouds by condensation processes and through capture by artificial obstacles. The main difficulty in using fog catchers is finding a design which gives accurate and representative readings of the horizontal precipitation within the ecosystem under study (Baynton, 1969). Another problem relating to the installation of the fog catcher is the selection of the level above the ground at which it is located (Ekern, 1964).

Table 2 shows, in a summarized form, the results of horizontal precipitation calculations with fog catchers in different cloud forests within the tropics. The values indicated within the table show a great variability at the absolute level, as well as the relative level compared to rainfall.

Absolute values vary between 325 mm/y and 941 mm/y; and the relative values between 7.2% and 158.5% of rainfall. Relative values for extremely rainy climates or seasons remain quite low (between 7.2% and 18% of rainfall equivalent).

On the other hand, it is important to note that during the dry seasons (Vogelmann, 1973) the relative values of horizontal precipitation are extremely high and can even exceed those of rain Juvik and Ekern (1978) mention a reading of 181.9 mm of horizontal precipitation in 3 dry months with only 14.5 mm of rainfall in a site located near a peak, which is the equivalent of 1,254% of rainfall. It is very probable that the relatively high quantity of water received through horizontal precipitation during the "dry" periods plays an important role in cloud forest ecology.

Juvik and Ekern's (1978) results additionally show that windward sites generally receive a great deal more horizontal precipitation than leeward ones, in absolute as well as relative values. However, in many cloud forests, where not only the slopes but also the peaks are swathed in clouds carried by winds, it is possible to observe what is known as the "spill-over effect" immediately on the leeward side of the crest where horizontal precipitation can reach very high levels. See data from Honulalai of Juvik and Ekern (1978) in table 2.

Table 2. Contribution of horizontal precipitation according to different studies undertaken in specific of the humid tropics

Author (Year)	Country; location	Surrounding vegetation	Specific conditions (mm/year)	Horizontal precipitation(% of rainfall)	Horizontal precipitation
Baynton (1969)	Puerto Rico; Luquillo Mountain (Pico del Oeste)	elfin cloud forest	1050masl; very rainy	325 390	7,2 8,6
Vogelmann (1973)	Mexico; Sierra Madre	oak cloud forest	1330masl; winy season
			(14 weeks) dry season (39 weeks) TOTAL 1,	941	18,0 85,5 31,3
			1898 mast; winy season (21 weeks) dry season (32 weeks) TOTAL	462	14,0 102,'9 22,3
Juvik and Ekern (1978)	Hawaii; Kuluni Camp Kulani Mauka Mauna Loa Honuaula Honulalai	Closed forest Alpine shrubs (lava)	windward: 1580 masl 2530 masl 3415miisl	770 706 186	30,0 68,0 43,4
		Pastureland Subalpine shrubs	leeward: 1905 masl. 2496 masl.	226 569	25,9 158,5
Caceres (1981)	Costa Rica Atlantic	Premontane rain forest	1300 masl. 550 slope	approx.	1 5,3

Caceres (1981) also reports, in addition to the readings of fog catchers, interesting values for net precipitation (throughfall and stem flow). Net precipitation varies between 82% and 99%, with an average of 92%, of the rain measured outside the forest. In a broad study carried out on the hydrological cycle of an Andean cloud forest, Steinhardt (1978) measured gross precipitation (outside the forest) and the net precipitation. The latter represents 90% of the gross precipitation. In both studies stem flow did not amount to even 1% of precipitation.

In comparison with other studies carried out on tropical forest interception, mentioned by Baumgartner and Brunig (1978), the readings cited by Steinhardt (1978) and Caceres (1981) for net precipitation are extremely high, which indicates a significant and effective contribution of horizontal precipitation to net precipitation.

Another study, not included in table 2, is that of Dohrenwend (n.d.) that determined, through the use of fog catchers, the contribution of horizontal precipitation to sub-alpine tropical vegetation resulting in approximately 20% of additional rainfall equivalent.

It is worthwhile mentioning here that solitary trees collect a great deal more horizontal precipitation per surface area than forests of the same species (Ekern, 1964; Vogelmann, 1973). According to Kammer (1974) this is due to the efficiency of the vegetation to collect and condense cloud moisture through exposure to wind. Merriam (1973), who studied horizontal interception with artificial leaves in controlled conditions, in a wind tunnel, concluded that:

• the quantity of water that enters and crosses the canopy is determined by the quantity of water held in the clouds, the height and arrangement of the forest canopy and the velocity and turbulence of the wind. '[`he latter, according to Lamb (1965), is increased by the forest and its physical characteristics (roughness parameter);
• the horizontal precipitation (fog drip) in turn depends on the total leaf surface, the special distribution of the leaves, and the physical properties of their surfacies.

Little is known of the effect of form and size of the leaves on horizontal interception, however, Went (1955) and Vogelmann (1973) consider that conifers are more efficient than broadleaved species.

Macroclimatic aspects. 'I'he relation.between macroclimate and cloud forests and their occurrence has been discussed in chapter 2. In general, macroclimatic parameters affecting cloud forests are:

• Grubb and whitmore (1966) suggest that the most important factor is the frequency of clouds. Lower montane rain forest and upper montane rain forest are associated with frequent and persistent clouds respectively;
• the structure of the troposphere, including the temperature profile, affect the level of cloud condensation (Flemming, 1971); the mass elevation effect influences the distribution and vertical temperature profile and thus cloud formation (Richards, 1952; Hastenrath, 1968). Grubb (1971, 1977) thinks that the occurrence of clouds plays an important role within the mass elevation effect ("Massenerhebungseffekt"); the direction and velocity of predominant winds (Kammer, 1974) together with the average atmospheric humidity (Kerfoot, 1968), are important factors in orographic cloud formation; the temporal distribution of rains and the presence of clouds (horizontal precipitation). Vogelmann's (1973) and Juvik and Ekern's (1978) measurements (see table 2) indicate that the effects of a dry season can be mitigated or compensated for by horizontal precipitation. This phenomenon could be an important ecologica' factor, in that Lauer (1952) showed that 1,000 mm of precipitation spread over twelve months has the same biological effect as 2,200 mm in nine months; the frequent cloud cover not only effects precipitation, but also other climatic elements as well as physiological processes. Budowski (1966) indicates that cloud cover represent an effective protection against radiation and large differences in temperature and relative humidity. Crubb and Whitmore (1966) and Baynton (1969) mentioned that the quantity of light can be reduced to a level at which assimilation is limited. Drewes and Drewes (1957) think that the "wet cloud forests" of the Eastern Andes are not primarily the result of an excessive precipitation but rather the lack of sunshine, resulting in a low evaporation rate and cool temperatures, and to the condensation of cloud moisture on plant surfaces; the diameter of water droplets contained in clouds strongly influences the condensation process and the deposition of water on vegetation surfaces and therefore the quantity of horizontal precipitation. Grunow (1960b) found that the most effective diameter of water droplets in the horizontal precipitation process is 8-14; another most important climatological phenomenon in tropical areas is the trade wind inversion which modifies the vertical temperature profile, thus influencing cloud formation (Riehl, 1954, 1979). The trade wind inversion is instable in that it is subjected to daily and annual variations. In addition, according to Dohrenwend (1972), the trade wind inversion that forms part of a Hadley cell, is situated at lower elevations in regions that are located far from the intertropical convergence zone (ITCZ), and higher in areas nearer the ITCZ. (see figure 8). This could intensify the mass elevation effect in certain areas, for example, the Caribbean as opposed to Central America. The areas in which trade wind inversion occurs are shown in figure 9;

Figure 8: Meconisms forming the trade wind inversion, according to Dohrenwend (1972)

Figure 9: Areas of trade wind inversion occurrence, occording to Dohronwend (1972)

Mesoclimatic aspects. With reference to the meso and topoclimatic scales, little information exists relating to tropical cloud forests. Huber (1976), in his research on the ecology of the cloud forest in Rancho Grande (Venezuela), divides it into three types according to structural differences:

• transition cloud forest with three storeys and trees that frequently extend beyond the canopy (approximately between 800 and 1,100 masl.
• true cloud forest with two storeys, with some trees extending beyond the canopy, palms abound and maximum of epiphyte syunsiae in quantity and diversity (approximately between 1,100 and 1,600 masl);
• upper level cloud forest, with a storey of dominating trees, a storey of dominated palms, and much fewer epiphytes than in the two other types (above 1,600 masl.

Huber (1976) came across these three types over a horizontal distance of two kilometres. It is likely that, apart from altitudinal differences, other climatic effects at the mesoclimatic level, especially the density and frequency of clouds and wind velocity modified by the topography, would contribute to these vegetational differences.

Troll (1968) mentioned two topoclimatic effects which could influence cloud forests:

• one refers to cloud formation due to circulation of air masses generated by the topography on the valleys of the Eastern Andes (see figure 10). These cloud banks condition cloud forests of "ceja de montana" and allow the "Yungas" forests to extend as a wide zone towards the interior of the mountains;
• another topoclimatic phenomenon described by Troll (1968) refers to the general vegetional limit which can be of particular importance to cloud forests. Comparing tropical and boreal zones, forest vegetation in the latter reaches its highest limits in mountain ridges; on the other hand, in the tropics trees find their most favourable conditions in valleys (see figure 11). Among the explanations given by Troll (1968) of this phenomenon, there are some which could hold true for cloud forests that frequently extend to extremely high elevations in mountain valleys (Hueck, 1978; Mann, 1968): daily temperature variations are generally less and air humidity is greater in the valleys than on the ridges. These two phenomena correspond with typical climatic characteristics of cloud forests and can therefore modify their distribution at the topoclimatic level.

In addition, topography can significantly modify wind action in certain mountain locations, increasing wind velocity and thus speeding up the exchange between atmosphere and vegetation. This can augment the quantity of water entering the ecosystem by means of horizontal precipitation, or strongly alter the structure of the forests conditioning elfin woodlands at the topoclimatic level (wind forests, according to Ashton et al., 1978).

Figure 10 Daytime local wind sydem in an eastern Andeon volley, according to Troll (1968) a) longitudinal section

Figure 10 Daytime local wind sydem in an eastern Andeon volley, according to Troll (1968) b) cross section

Figure 11 The effect of topoclimatic conditions on the altitude of the timber line in mountains of the tropical and boreal belt, occording to troll (1968) - Tropical Zone and Boreal Zone

Microclimatic aspects. Richards (1952) gives an extensive summary of tropical rain forest microclimate in general. However only a few limited studies exist of microclimates within cloud forests.

Beebe and Crane (1947), as well as Huber (19763, included some simple microclimatic measurements (mainly temperature, humidity and light) in their ecological studies of the cloud forest at Rancho Grande (Venezuela). In addition Huber (1976, 1978) determined the light compensation point of 54 species resulting in a very large variability. Baynton (1968, 1969) investigated the microclimate of a dwarf cloud forest in Puerto Rico, including the establishment of wind profiles. Lotschert (1959) included evaporation measurements within his microclimatic studies of a cloud forest in El Salvador.

Apart from the general characteristics of microclimate within humid tropical forests (Richards, 1952) it is possible to summarize the microclimatic properties of cloud forests as follows:

The principal microclimatic parameter affecting cloud forests is the high relative humidity of the air in combination with horizontal precipitation. These two elements, very often associated with rather low temperatures, keep cloud forests permanently humid. This in turn facilitates the presence of epiphytes (mosses and lichens) which are able to keep the microclimate humid even when, at the macroclimatic level, the relative humidity has dropped (Leigh, 1975; Grubb, 1977; Tanner, 1980b). In elfin woodlands the reduction in transpiration below the canopy, due to the abundance of mosses, can result in a complete obstruction of development of undergrowth or any other vegetation on the soil away from the mosses (Leigh, 1975).

Wind generally favours transpiration by reducing the external resistance to transpiration. However, extremely humid air, particularly that charged with water droplets, can block transpiration with a permanent layer of water on the leaves. This effect and its ecological importance within cloud forests has been discussed in chapter 3 (Other Language Terms). The hydrological aspects of reduced transpiration in cloud forests at watershed level is discussed below. Crubb (1977) indicated that the high humidity favourized the "invasion of lichens and bryophytes" and, referring to Berrie and Eze (1975), mentions the damage which can be caused by such invasions; not only by covering and obscuring the leaves, but also through the destructive effects on the cuticula. Crubb (1977) considers that "infections" of these epiphyllic organisms, conditioned by the extremely humid microclimatic environment, cause the most serious damage to the vegetation. It therefore appears strange that drip tips* are lacking in cloud forests. Crubb (1977) supposes that these are only effective in climatic conditions when heavy rain (or storms) are interspersed with periods of sunshine, but are ineffective in permanently cloudy conditions. Ellenberg (1975), who also mentions the lack of drip tips in cloud forests, on the other hand, actually questions the teleological interpretation of this phenomenon.

Another relevant microclimatic phenomenon that has attracted the attention of various authors is the presence of xeromorphism within cloud forests in the humid tropics. Walter (1979) considered that the leaves on the trees, even in the most humid tropical zones, were exposed to solar radiation for several hours thus resulting in a heating up of the leaves by 10 K above air temperature. Ellenberg (1959) mentions xeromorphism of epiphytes, giving the same explanation as Walter. Leigh (1975), on the other hand, considers that there is still no satisfactory explanation of the phenomenon that, according to him, assumes a considerable importance in the formation of a thick layer of undecomposed organic matter frequently found in cloud forests. Crubb (1977) rejects the climatic explanation of xeromorphism in cloud forests; for him physiological factors and nutritional effects determine the xeromorphism of leaves in cloud forests.

The formation of xeromorphic structures in cloud forests can also be interpreted as a protective mechanism for plant surfaces against the chemical impacts of horizontal precipitation which, according to Lovett, Reiners and Olson (1982) have different chemical properties from the rainfall, tending to be much more acid (Falconer and Falconer, 1980; Schrimpff et al., 1984). According to Falconer and Falconer (1980), the acidity of cloud moisture droplets is more pronounced in humid tropical air masses with dew point > 15°C.

It is worth mentioning here that a great deal of microclimatic research remains to be carried out in different types of cloud forests to increase knowledge of its ecological and hydrological importance. To date only dwarf cloud forests have been the object of more detailed microclimatic studies and considerations in that their unusual appearance has attracted the attention of various scientists (see chapter 3, Terminology -Elfin Woodlands).