Introduction
Tropical montane forests contain enormous epiphyte species richness (Pócs Reference Pócs and Newmark1991; Agrawala et al. Reference Agrawala, Moehner, Hemp, Van, Hitz, Smith, Meena, Mwakifwamba, Hyera and Mwaipopo2003; Hemp Reference Hemp2005). Most studies have described the forest vegetation of East African mountains based on trees, and also pteridophytes (Hemp Reference Hemp2002) and bryophytes (Pócs Reference Pócs and Newmark1991; Lovett & Pócs Reference Lovett and Pócs1993; Mattila & Koponen Reference Mattila and Koponen1999) including liverworts (Pócs Reference Pócs and Newmark1991; Lovett & Pócs Reference Lovett and Pócs1993). Only a few studies have focused on lichen species (Swinscow & Krog Reference Swinscow and Krog1988; Sipman & Harris Reference Sipman, Harris, Lieth and Werger1989; Pócs Reference Pócs and Newmark1991; Kirika et al. Reference Kirika, Ndiritu, Mugambi, Newton and Lumbsch2018). Lichens are mutualistic symbiotic organisms composed of a fungal partner (the mycobiont), and one or more photosynthetic partners (the photobiont), which are either a green alga or a cyanobacterium (Galloway Reference Galloway1992; Scheidegger & Goward Reference Scheidegger, Goward, Nimis, Scheidegger and Wolseley2002; Werth & Scheidegger Reference Werth and Scheidegger2012; Dal Grande et al. Reference Dal Grande, Beck, Cornejo, Singh, Cheenacharoen, Nelsen and Scheidegger2014; Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin, Werth, Cheenacharoen and Scheidegger2014b). The lichen symbiosis is one of the most successful symbioses known in nature, being found in all parts of the world, in a vast spectrum of microhabitats and microclimates (Galloway Reference Galloway1992). In most terrestrial ecosystems lichens can survive in situations where higher plants cannot grow (Liska et al. Reference Liska, Detinsky and Palice1996). Nevertheless, lichens are among the most neglected taxa in tropical rain forest studies. In addition, lichenological knowledge throughout the African region remains inadequate (Sipman & Harris Reference Sipman, Harris, Lieth and Werger1989). A study has shown that the number of lichenologists in Tanzania has been either low or practically non-existent in comparison with the number of scientists specializing in vascular plants, and data on the distribution of lichen species was absent or unpublished (Pócs Reference Pócs and Newmark1991).
This study investigates the ecology and local population sizes (frequency of occurrence) of the epiphytic macrolichen Lobaria pulmonaria (L.) Hoffm. in tropical montane forests on Mt Kilimanjaro and Mt Meru in Tanzania. This species, commonly known as tree lungwort, is an epiphytic foliose tripartite macrolichen species containing fungal (Ascomycetes), green-algal (Symbiochloris reticulata; Škaloud et al. Reference Škaloud, Friedl, Hallmann, Beck and Dal Grande2016) and cyanobacterial (Nostoc sp.) partners (Jordan Reference Jordan1970; Scheidegger et al. Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998; Snäll et al. Reference Snäll, Pennanen, Kivistö and Hanski2005; Coxson & Stevenson Reference Coxson and Stevenson2007a; Dal Grande et al. Reference Dal Grande, Widmer, Beck and Scheidegger2010; Larsson & Gauslaa Reference Larsson and Gauslaa2011; Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin, Werth, Cheenacharoen and Scheidegger2014b). Due to its wide distribution in humid parts of Europe, Asia, North America and Africa (Liska et al. Reference Liska, Detinsky and Palice1996; Scheidegger et al. Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998; Zoller et al. Reference Zoller, Lutzoni and Scheidegger1999; Walser et al. Reference Walser, Zoller, Büchler and Scheidegger2001), it is used as an indicator species for rapid assessment of the conservation importance of forests (Nascimbene et al. Reference Nascimbene, Marini and Nimis2007; Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin and Scheidegger2014a) and an important model species for studies on the conservation biology of epiphytic lichens (Walser et al. Reference Walser, Zoller, Büchler and Scheidegger2001, Reference Walser, Sperisen, Soliva and Scheidegger2003; Scheidegger & Werth Reference Scheidegger and Werth2009; Jüriado et al. Reference Jüriado, Liira, Csencsics, Widmer, Adolf, Kohv and Scheidegger2011). Not only is it among the most productive lichens, which provide N-fertilization to forest ecosystems (Campbell & Fredeen Reference Campbell and Fredeen2004; Ellis Reference Ellis2012; Gauslaa & Goward Reference Gauslaa and Goward2012), but also an indicator species for primeval forests (Liska et al. Reference Liska, Detinsky and Palice1996; Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin and Scheidegger2014a) and often used as a flagship species in practical conservation because it is easily recognized by foresters and naturalists (Scheidegger et al. Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998).
Studies have shown that some environmental variables such as forest type and altitude (Liska et al. Reference Liska, Detinsky and Palice1996; Gu et al. Reference Gu, Kuusinen, Konttinen and Hanski2001), light regimes (Scheidegger Reference Scheidegger1995; Gauslaa & Solhaug Reference Gauslaa and Solhaug2000; Mackenzie et al. Reference MacKenzie, MacDonald, Dubois and Campbell2001), and habitat moisture (Liska et al. Reference Liska, Detinsky and Palice1996; Gauslaa Reference Gauslaa2014) may affect the distribution of L. pulmonaria (Liska et al. Reference Liska, Detinsky and Palice1996; Mackenzie et al. Reference MacKenzie, MacDonald, Dubois and Campbell2001). Many studies have also shown that L. pulmonaria has a low effective dispersal owing to high juvenile mortality (Walser Reference Walser2004; Öckinger et al. Reference Öckinger, Niklasson and Nilsson2005; Werth et al. Reference Werth, Wagner, Gugerli, Holderegger, Csencsics, Kalwij and Scheidegger2006a; Otàlora et al. Reference Otàlora, Martinez, Belinchon, Widmer, Aragon, Escudero and Scheidegger2011). It therefore occupies old-growth forest and occurs most abundantly on large trees (Öckinger et al. Reference Öckinger, Niklasson and Nilsson2005; Snäll et al. Reference Snäll, Pennanen, Kivistö and Hanski2005), where it is restricted to bark surfaces with relatively high pH between 5.0 and 6.0 (Gauslaa Reference Gauslaa1985, Reference Gauslaa1995; Rose Reference Rose1988; Scheidegger Reference Scheidegger1995; Kermit & Gauslaa Reference Kermit and Gauslaa2001; Carlsson & Nilsson Reference Carlsson and Nilsson2009). In recent decades, several studies have documented the loss of populations of L. pulmonaria because of air pollution (Rose Reference Rose1988; Scheidegger Reference Scheidegger1995; Gu et al. Reference Gu, Kuusinen, Konttinen and Hanski2001) and forest management (Gu et al. Reference Gu, Kuusinen, Konttinen and Hanski2001), and the species is considered as regionally rare and threatened (Nascimbene et al. Reference Nascimbene, Marini and Nimis2007; Carlsson & Nilsson Reference Carlsson and Nilsson2009; Catalano et al. Reference Catalano, Mingo, Migliozzi, Sgambato and Aprile2010; Dal Grande et al. Reference Dal Grande, Widmer, Beck and Scheidegger2010; Larsson & Gauslaa Reference Larsson and Gauslaa2011). As a result, this has led to several studies researching the conservation biology and ecology of the species (Gu et al. Reference Gu, Kuusinen, Konttinen and Hanski2001; Gauslaa et al. Reference Gauslaa, Lie, Solhaug and Ohlson2006; Nascimbene et al. Reference Nascimbene, Marini and Nimis2007). Furthermore, L. pulmonaria is widely used to evaluate the spatio-temporal continuity of forest cover (Rose Reference Rose, Brown, Hawksworth and Bailey1976) and to assess environmental quality in areas of high biogeographical interest (Catalano et al. Reference Catalano, Mingo, Migliozzi, Sgambato and Aprile2010). However, nothing is known about this species’ distribution and ecology in the tropical montane forests of Mt Meru and Mt Kilimanjaro, and a detailed knowledge of the distribution and habitat requirements of L. pulmonaria in the tropics remains essential to plan conservation action and develop comprehensive policies for this threatened species and others with a similar ecology.
The present study examines the target species, L. pulmonaria, in terms of population sizes in relation to host tree variables in tropical montane forest landscapes. We hypothesized that: 1) L. pulmonaria is associated with large host trees; 2) L. pulmonaria forms larger thalli on large and old host trees than on young host trees; 3) factors associated with host trees’ microhabitats explain the occurrence of L. pulmonaria; 4) L. pulmonaria does not occur randomly but only on specific tree species.
Material and Methods
Study area
Mt Kilimanjaro is Africa's highest mountain and was given the status of a Biosphere Reserve and World Heritage Site in 1989 (Noe Reference Noe2014) It is located 300 km south of the equator in Tanzania on the border with Kenya, between 2°45′–3°25′S and 37°0ʹ–37°43ʹE (Hemp Reference Hemp2002, Reference Hemp2006). It is 90 km wide from north-west to south-east (Hemp Reference Hemp2002) and represents an eroded relic of an ancient volcano with three peaks, Kibo, Mawenzi and Shira, that reach altitudes of 5895 m, 5149 m and 3962 m a.s.l., respectively (Agrawala et al. Reference Agrawala, Moehner, Hemp, Van, Hitz, Smith, Meena, Mwakifwamba, Hyera and Mwaipopo2003; Hemp Reference Hemp2005). Arusha National Park is the core region of Mt Meru located north-east of Arusha town in the Arumeru District of northern Tanzania, at 3°15ʹS, 36°45ʹE (Martinoli et al. Reference Martinoli, Preatoni, Galanti, Codipietro, Kilewo, Fernandes, Wauters and Tosi2006; Giliba et al. Reference Giliba, Mafuru, Paul, Kayombo, Kashindye, Chirenje and Musamba2011). It is the African continent's fifth highest mountain (4566 m a.s.l.) and its topography is that of a young volcano of Pleistocene origin (Instituto Oikos 2011).
Climate
Our study focused on the southern part of Mt Kilimanjaro which is characterized as having a typical equatorial climate. The northern slopes, on the lee side of the mountain, receive much less annual rainfall than the southern slopes (Hemp Reference Hemp2002, Reference Hemp2005). The study area has two distinct rainy seasons: the long rains from March to May, and the short rains around November. The annual precipitation reaches its maximum c. 3000 mm in the mid-montane zone, between 1800 m and 2400 m (Agrawala et al. Reference Agrawala, Moehner, Hemp, Van, Hitz, Smith, Meena, Mwakifwamba, Hyera and Mwaipopo2003; Hemp Reference Hemp2005). In the alpine zone, however, the precipitation decreases to c. 200 mm (Hemp Reference Hemp2002). The mean annual temperature decreases linearly upslope, with a lapse rate of 0.56 °C per 100 m (Hemp Reference Hemp2005) starting at 23.41 °C at the foothills in Moshi (813 m) and decreasing to −7.11 °C at the top of Kibo (Agrawala et al. Reference Agrawala, Moehner, Hemp, Van, Hitz, Smith, Meena, Mwakifwamba, Hyera and Mwaipopo2003). The climate on Mt Meru is characterized by two distinct seasonal weather patterns: short rains in November and December, and long rains from mid-March to late May. The southern slopes of Mt Meru receive up to 2000 mm of rainfall per year, with an annual maximum in the montane belt of c. 2200 mm. The annual mean temperature is 18.68 °C and the hottest season is between January and February with the temperatures sometimes exceeding 25 °C, while the cold season is from June to August with the temperature at midday just below 16 °C. Furthermore, on the highest parts of Mt Meru, temperatures are lower and frost occurs at night during the cold season. Due to their high elevation, both Mt Kilimanjaro and Mt Meru have a greater orographic influence on the climate and can be considered wet ‘islands’ in a dry season (Instituto Oikos 2011).
Vegetation
According to Hemp (Reference Hemp2005), several bioclimatic belts can be distinguished along the slopes of Mt Kilimanjaro. A dry and hot colline savannah zone surrounds the mountain base between 700 m and 1000 m (mostly farmland, and some intact savannah grassland). The sub-montane belt between 1000 m and 1800 m has been converted to coffee-banana plantations, with montane forests covering an area of c. 1000 km2 on Mt Kilimanjaro (Hemp Reference Hemp2005). In the western parts of the mountain, the comparatively dry submontane forest below 1600 m is dominated by Olea europaea ssp. africana, Croton megalocarpus, Cassipourea malosana and Diospyros abyssinica. Above 1600 m the most important tree is the camphor-tree Ocotea usambarensis. In a lower altitudinal zone, it occurs commonly with Agauria salicifolia and Macaranga kilimandscharica. In an upper zone it is associated with Podocarpus latifolius (Hemp Reference Hemp2002). Similarly, on the drier northern slope, the lower montane zone is dominated by Croton-Calodendrum forests. Mid-altitudes there are dominated by Cassipourea forests, whereas Juniperus procera characterizes the higher altitudes (Hemp Reference Hemp2005). On the southern slope of the mountain, the mature forest reaches a canopy height of up to 40 m or more, is dominated by the camphor-tree Ocotea usambarensis and characterized by a high abundance of epiphytes (Schrumpf Reference Schrumpf2004; Schrumpf et al. Reference Schrumpf, Zech, Lehmann and Lyaruu2006). The cloud forest zone is dominated by Podocarpus latifolius, Hagenia abyssinica and Erica excelsa; and above 2800 m Erica excelsa is the dominant tree species (Hemp Reference Hemp2005). Above c. 3100 m, these forests have been replaced by Erica bush in recent decades; here the species E. arborea and E. trimera, Protea kilimandscharica and Euryops dacrydioides cover large areas of the subalpine zone (Hemp & Bayreuth Reference Hemp and Bayreuth2001). Around 3900 m elevation, the Erica bush grades into Helichrysum cushion vegetation with H. newii and H. citrispinum reaching c. 4500 m, while higher altitudes are poorly vegetated (Hemp Reference Hemp2002, Reference Hemp2005). Moreover, a high number of rare and endemic plants occur on Mt Kilimanjaro, including Dendrosenecio johnstonii, Diospyros kilimandscharica, Disperis kerstenii, Euphorbia quinquecostata, Euryops dacrydioides Gerrardinia eylesiana, Helichrysum meyeri-johannis (decimated along climbing trails), Impatiens kilimanjari, I. digitata, I. volkensii, Lobelia deckenii, Psychotria petiginosa, Rubus volkensii and Vepris arushensis. Eleven species of bryophyte are endemic to Mt Kilimanjaro, including Colura berghenii, Pocsiella hydrogonioides and Zygodon robustus (Lovett & Pócs Reference Lovett and Pócs1993).
According to Instituto Oikos (2011), the vegetation of Mt Meru can be divided into three main zones: the mountain forest zone (evergreen forest) is dominated by multi-stratified closed evergreen trees with heights up to 30 m or more. It includes the dry montane forests, on the eastern, north-western and northern slopes, between 1500–2600 m elevation, with the threatened Juniperus procera, Olea europaea ssp. africana and Podocarpus falcatus; the moist montane forest, on the eastern and western slopes between 1500 and 2400 m elevation, with a great number of broad-leaved hardwood trees; and the upper montane forest between 2300–3000 m elevation, with the bamboo thickets (Arundinaria alpina) and Hagenia abyssinica forests. The Sub-Afroalpine heath zone (ericaceous zone), between 3000–3600 m elevation, is dominated by arborescent or shrubby species of Erica and Philippia. The Afroalpine zone, up to the summit at 4560 m, has poor floral species diversity, mainly consisting of open steppe-like communities, dominated by two species of Pentaschistis grass.
Sampling design
On Mt Kilimanjaro, Lobaria pulmonaria thalli were collected from Marangu, Mweka, Umbwe and Machame between 2700 m and 3100 m elevation (Fig. 1). On Mt Meru, thalli were collected from gradients below the Miriakamba hut, above the Miriakamba hut and below the Saddle hut, between 2100 m and 3300 m elevation (Fig. 1). For each route sampled along the altitudinal gradients, host trees greater than 5 cm in diameter at breast height were searched for L. pulmonaria (Wagner et al. Reference Wagner, Holderegger, Werth, Gugerli, Hoebee and Scheidegger2005; Öckinger & Nilsson Reference Öckinger and Nilsson2010). In total, 1055 thalli of L. pulmonaria were sampled and their diameter measured (longest axis), from 389 host trees in forest stands organized as 13 (c. 1-ha) plots (Supplementary Material Table S1, available online). Additionally, a total of 4532 thalli of L. pulmonaria were counted on trunks from these sampled host trees up to 5 m above ground level (Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin, Werth, Cheenacharoen and Scheidegger2014b). We also recorded other phorophyte variables such as phorophyte species, phorophyte height which was measured with a Blume-Leiss altimeter, phorophyte size (circumference of each sampled phorophyte was measured at breast height, c. 1.3 m above ground level), bark texture (Öckinger et al. Reference Öckinger, Niklasson and Nilsson2005; Öckinger & Nilsson Reference Öckinger and Nilsson2010) and the diameter of each thallus was measured using a metric ruler. Host tree trunk inclination was estimated and if the trunk was straight up to 2 m height, the host tree was evaluated as straight (Mežaka et al. Reference Mežaka, Brūmelis and Piterāns2008). Host tree bark samples (0.5 g) were collected from beneath L. pulmonaria thalli for pH measurements in the laboratory using standard procedures (Gauslaa Reference Gauslaa1995). Each host tree was geo-referenced by a hand-held Garmin GPS.
A map of Mt Kilimanjaro and Mt Meru in Tanzania. Insets show the study sites with a scale provided and annotations which relate to the sampled plots and the phorophytes species (green circles). For further details see Table 3.
Statistical analyses
Data were analyzed using R v. 4.1.0 (R Core Team 2021). We performed two types of analyses. Firstly, the statistical tests were treated in a preliminary and descriptive manner to discriminate between relevant and uninteresting results and determine the association of L. pulmonaria with host trees among the study sites and regions. Secondly, linear mixed-effects models (LME) were used to determine the relationship of host tree variables, and, as response variables i) the diameter of L. pulmonaria thalli sampled per host tree, ii) the number of L. pulmonaria thalli counted per host tree. The analyses were performed with two datasets. The dataset with 4532 L. pulmonaria thalli that were counted from the trunks of respective host trees (n = 389) (combined over the altitudinal gradients on Mt Kilimanjaro and Mt Meru), and another dataset with 1055 L. pulmonaria thalli that were sampled from respective host trees (combined over the plots on Mt Meru and Mt Kilimanjaro). Relationships between host tree variables and the number of L. pulmonaria thalli were tested with LME fitted by restricted maximum likelihood (REML). The number and size of thalli were log-transformed to obtain the normal distribution of residuals. The model selection was then carried out using an automated stepwise selection procedure based on Akaike's information criterion (AIC) to find an optimal model according to predictive power and to avoid overparameterization (Shao Reference Shao1997; Jüriado et al. Reference Jüriado, Karu and Liira2012). The ‘stepAIC’ function was performed in the MASS package (Venables & Ripley Reference Venables and Ripley2002). To check for heteroscedasticity (Kormann et al. Reference Kormann, Rösch, Batàry, Tscharntke, Orci, Samu and Scherber2015), models with and without a variance function were implemented using restricted maximum likelihood. AIC values indicated that the model without a variance function considerably improved the model fit for the number of thalli in relation to the host tree. The ggplot2 package (Wickham Reference Wickham2009) was used to display the graphical distribution of L. pulmonaria thalli number and size with host trees among the study sites and regions.
Results
Distribution of Lobaria pulmonaria with host tree size
In the first data set, analyzed by the linear-mixed effects models using the ‘stepAIC’ function, two host tree variables (i.e. circumference (Fig. 2A) with AIC = 300.30 and height (Fig. 2B) with AIC = 305.94) were the best predictors of the number of thalli counted per trunk. Whereas in the second data set, the analysis retained trunk shape (Fig. 3A) with AIC = 2225.3, bark texture (Fig. 3B) with AIC = 2225.0, trunk circumference (Fig. 4A) with AIC = 2225.2 and bark pH (Fig. 4B) with AIC = 2229.2 as the best predictors for maximum thallus size of L. pulmonaria. The results for the first data set (Table 1) indicate a significant effect of host tree height on the distribution of L. pulmonaria species among the study sites and across the regions (DF = 375, P < 0.0001). Also, there is weak evidence that trunk circumference may influence L. pulmonaria species in occupying the host trees (P = 0.0873). The results of the second data set (Table 2) indicate a significant effect of trunk circumference (P = 0.0299) on Lobaria pulmonaria thallus size among the study sites and across the regions. Also, there is strong evidence that host tree bark pH (P = 0.0123) influences the L. pulmonaria thallus size but weak evidence was found for the influence of trunk shape (P = 0.1896) and bark texture (P = 0.1881) on L. pulmonaria thallus size among the study sites and across the regions.
Relationship between the number of thalli of Lobaria pulmonaria counted per trunk and phorophyte circumference (A) and height (B), in the forests of Mt Kilimanjaro and Mt Meru in Tanzania. These two host tree variables were the best predictors of the number of thalli counted per trunk in the data set analyzed by the linear-mixed effects models.
Relationship between Lobaria pulmonaria thallus diameter, and phorophyte trunk shape (A) and bark texture (B) in the forests of Mt Kilimanjaro and Mt Meru in Tanzania. In this analysis linear mixed-effects models were used to determine the relationship of the host tree variables, with the diameter of L. pulmonaria thalli sampled per host tree as a response variable.
Relationship between Lobaria pulmonaria thallus diameter, and phorophyte circumference (A) and bark pH (B) in the forests of Mt Kilimanjaro and Mt Meru in Tanzania. In this analysis linear mixed-effects models were used to determine the relationship of the host tree variables, with the diameter of L. pulmonaria thalli sampled per host tree as a response variable.
Sequential analysis of variance for the effects of height and circumference on the number of Lobaria pulmonaria thalli per phorophyte (linear mixed effect model).
Sequential analysis of variance for the effects of phorophyte bark texture, circumference, trunk shape and bark pH on the size of Lobaria pulmonaria thalli per phorophyte (linear mixed effect model).
Distribution of Lobaria pulmonaria with tree species
In total there were 301 host trees associated with the 10 populations of L. pulmonaria recorded in the forests of Mt Kilimanjaro and 88 host trees among three populations in the forests of Mt Meru (Table 3, Fig. 5A & B). Among the host tree species that were surveyed in the forests of Mt Kilimanjaro, there were five which appeared to host a greater number of L. pulmonaria thalli, including Ilex mitis with 0.26 frequency of occurrence, Podocarpus latifolius (0.18), Senecio sp. (0.16), Prunus africana (0.11) and Rapenea melanophloeos (0.09). On Mt Meru, as shown in Table 3 and Fig. 5B, the occurrence of L. pulmonaria was overall less frequent; however, among the host trees that were surveyed, there were also five tree species which appeared to host a relatively high number of L. pulmonaria thalli, including Bersama abyssinica with 0.33 frequency of occurrence, Hypericum revolutum (0.27), Hagenia abyssinica (0.14), Rapenea melanophloeos (0.10) and Olea capensis (0.07). It was found that there were more host tree species associated with L. pulmonaria in the forests of Mt Kilimanjaro, including Catha edulis, Erica arborea, E. excelsa, Hypericum revolutum, Ilex mitis, Macaranga capensis, Maytenus acuminata, Ocotea usambarensis, Podocarpus latifolius, Prunus africana, Rapenea melanophloeos, Rhamnus prinoides, Senecio subsessilis and Senecio sp., than on Mt Meru which had fewer host tree species including Bersama abyssinica, Catha edulis, Clematis sp. (climber), Hagenia abyssinica, Hypericum revolutum, Mystroxylon aethiopicum, Olea capensis, Podocarpus latifolius, Prunus africana and Rapenea melanophloeos. Among the 14 host tree species where L. pulmonaria was recorded on Mt Kilimanjaro, thalli were more frequent on both small to medium sized and large old-growth host trees of Ilex mitis and Podocarpus latifolius than on any other host tree. While the results from Mt Meru indicated that L. pulmonaria thalli were more frequent on small to medium sized Bersama abyssinica than other host trees. Furthermore, the information given in Supplementary Material Table S1 (available online) and Table 3 indicated that L. pulmonaria was unevenly distributed elevationally among the sub-alpine to montane forest altitudinal gradients of Mt Kilimanjaro and Mt Meru.
Frequency of occurrence of phorophytes with Lobaria pulmonaria in the 13 sampled populations from the forests of Mt Kilimanjaro (A) and Mt Meru (B).
Frequency of phorophyte species recorded with Lobaria pulmonaria from the 13 sampled populations (plots) on Mt Kilimanjaro and Mt Meru in Tanzania.
n = number of host trees recorded
pH range = pH of sampled bark of host trees recorded
Np = number of phorophytes recorded with Lobaria pulmonaria
Ns = number of phorophytes sampled per study site
Discussion
Distribution of Lobaria pulmonaria with tree size
In this study, data collected from high elevation tropical forests, between 2500 m and 3200 m elevation on Mt Kilimanjaro, have indicated that L. pulmonaria occurs most frequently on Ilex mitis and Podocarpus latifolius (Fig. 5A). Also, it often occurred on small sized host trees such as Senecio sp. However, Macaranga capensis is one of the largest old-growth host trees and an important habitat of L. pulmonaria species, though was present only in one locality (MR4). These findings for Mt Kilimanjaro differ from the observations in the forest of Mt Meru, where L. pulmonaria occurred most frequently on Bersama abyssinica and Hypericum revolutum (shrub) (Fig. 5B). These host trees were small to medium sized in terms of height and circumference. Also, Hagenia abyssinica, a large old-growth host tree and a supposed important habitat of L. pulmonaria species on Mt Meru, had a few individual thalli which appeared to occur in only a small elevational range between 2570 m and 3220 m above Miriakamba hut. Comparatively, Scheidegger et al. (Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998) indicated that colonization by L. pulmonaria significantly increased with a tree's circumference, and that small trees with a circumference less than 47 cm were not colonised with L. pulmonaria. On the contrary, the results of this project have shown that, in tropical mountain forests, L. pulmonaria formed a large number of thalli even on small sized, young host trees with a circumference of c. 20 cm, with larger thalli on host trees with a circumference between 50 cm and 100 cm. Moreover, large thalli of L. pulmonaria were present, but only on a few thicker, old-growth host trees with a circumference over 300 cm. A possible explanation is that lichen fragments falling off nearby large old-growth hardwood host trees continue to grow on branches of small sized host trees in their vicinity. It would be interesting to conduct another study on the effect of distance between putative old-growth and young host trees in relation to the number and size of L. pulmonaria thallus per host tree.
Previous literature has demonstrated that old-growth forests are key reservoirs of epiphytic lichen species (Neitlich & McCune Reference Neitlich and McCune1997; Scheidegger et al. Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998; Sillett et al. Reference Sillett, McCune, Peck, Rambo and Ruchty2000; Bosch et al. Reference Bosch, Prati, Hessenmöller, Schulze and Fischer2013). Furthermore, cyanolichens and tripartite lichens such as L. pulmonaria are often categorized as old growth-dependent species (Muir et al. Reference Muir, Shirazi and Patrie1997; Öckinger et al. Reference Öckinger, Niklasson and Nilsson2005; Snäll et al. Reference Snäll, Pennanen, Kivistö and Hanski2005; Coxson & Stevenson Reference Coxson and Stevenson2007b). In wider relation to this, Lie et al. (Reference Lie, Arup, Grytnes and Ohlson2009) has shown that there is a relationship between tree size and lichen epiphytic diversity patterns, while Nadyeina et al. (Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin and Scheidegger2014a) and Liska et al. (Reference Liska, Detinsky and Palice1996) have also shown that L. pulmonaria prefers old-growth forests. Gauslaa (Reference Gauslaa1995) and Jüriado & Liira (Reference Jüriado and Liira2010) have discussed the characteristic feature of the species which is considered to be a representative of a climax lichen community on mature hardwood trees in old forests. Meanwhile, other studies (Gu et al. Reference Gu, Kuusinen, Konttinen and Hanski2001; Kalwij et al. Reference Kalwij, Wagner and Scheidegger2005; Öckinger et al. Reference Öckinger, Niklasson and Nilsson2005; Edman et al. Reference Edman, Eriksson and Villard2008) have clearly shown that in a forested landscape, the occurrence of L. pulmonaria within a stand is dependent on the diameter of host trees. Under optimal climate and habitat conditions where L. pulmonaria is abundant, the species can grow on various trees, even those with small diameters (Carlsson & Nilsson Reference Carlsson and Nilsson2009; Jüriado et al. Reference Jüriado, Karu and Liira2012). Edman et al. (Reference Edman, Eriksson and Villard2008) have reported a strong relationship between tree size and the abundance of lichens, indicating the importance of large trees in forest ecosystems. Therefore, our results concur with most of the above studies, while suggesting that there is a strong increase in the number of L. pulmonaria thalli with increasing host tree size.
Furthermore, our results indicate that many trees, regardless of their size, being small, medium sized to large old-growth trees, were potential habitats for a considerable number of L. pulmonaria thalli, including such shrubby host trees as Hypericum revolutum. Although the results have indicated that L. pulmonaria thalli were largely associated with young host trees, these findings do not invalidate suggestions that protecting old-growth host trees is of paramount importance for preserving biodiversity (Muir et al. Reference Muir, Shirazi and Patrie1997). In essence, the key challenge of relying on lichen communities associated with young host trees is that after catastrophic disturbance such as frequent fires, which are often re-occurring on Mt Kilimanjaro (Hemp Reference Hemp2001; personal field observation) and Mt Meru (personal field observation), the rate of recovery of epiphytes may depend on old-growth remnants beyond the reach of fire as propagule sources for recovery into in young forests (Nascimbene et al. Reference Nascimbene, Marini and Nimis2007).
Distribution of Lobaria pulmonaria with the type of host tree species
Our survey along altitudinal gradients in the tropical montane forests of Mt Kilimanjaro and Mt Meru, indicated that L. pulmonaria thalli were unevenly distributed among the sub-alpine and montane forests and we found that the species occurred in patches along the hills and valleys (observational data, unpublished). Similar findings were reported by McCune et al. (Reference McCune, Derr, Muir, Shirazi, Sillett and Daly1996) who observed the occurrence of L. pulmonaria along valleys, but also on hills (Rose Reference Rose1988; Liska et al. Reference Liska, Detinsky and Palice1996) and otherwise in patches (Nascimbene et al. Reference Nascimbene, Marini and Nimis2007). Furthermore, host tree distribution can cause patterns to emerge. It was observed that, in the mountains of Kilimanjaro, L. pulmonaria most frequently occurred on Hypericum revolutum (shrub), Ilex mitis, Podocarpus latifolius and Rapenea melanophloeos host trees, and less frequently on others including Catha edulis, Erica arborea, E. excelsa, Macaranga capensis, Maytenus acuminate, Ocotea usambarensis, Prunus africana, Rhamnus prinoides, Senecio subsessilis and Senecio sp. On Mt Meru, L. pulmonaria was mainly growing on Bersama abyssinica and Hypericum revolutum host trees and less frequently on Catha edulis, Clematis sp. (climber), Hagenia abyssinica, Mystroxylon aethiopicum, Olea capensis, Podocarpus latifolius, Prunus africana and Rapenea melanophloeos. In some instances, host tree species such as Hypericum revolutum and Rapenea melanophloeos were important habitats on both of the mountains while other host trees species such as Bersama abyssinica, Hagenia abyssinica and Ilex mitis were a good substratum host on one mountain only. The distribution of epiphytic lichens has been described at different spatial scales ranging from the single tree scale, to the forest stand and landscape scales, up to the regional scale (Lie et al. Reference Lie, Arup, Grytnes and Ohlson2009). Many studies combined indicate that L. pulmonaria does not have a specific preference for certain type of forests; for example, in European countries L. pulmonaria is found in different forest types, the most important being beech forests (Nadyeina et al. Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin and Scheidegger2014a), coniferous montane forests, chestnut forests and oak forests (Nascimbene et al. Reference Nascimbene, Brunialti, Ravera, Frati and Caniglia2010). In Finland, Gu et al. (Reference Gu, Kuusinen, Konttinen and Hanski2001) and Nadyeina et al. (Reference Nadyeina, Dymytrova, Naumovych, Postoyalkin and Scheidegger2014a) have reported that L. pulmonaria occurs throughout the country and has greatly declined but is not yet considered as a threatened species because it is still fairly common in the remaining eastern-central old-growth forests, especially on Populus tremula, Salix caprea and Sorbus aucuparia. In humid forests, L. pulmonaria has also been found on other tree species such as the lower branches of old spruce trees (Picea abies) and on vertical cliffs. The studies carried out by Walser (Reference Walser2004), Werth et al. (Reference Werth, Wagner, Holderegger, Kalwij and Scheidegger2006b) together with Werth & Scheidegger (Reference Werth and Scheidegger2012) have shown that the distribution of L. pulmonaria along the pasture-woodland landscape of the Jura Mountains in Switzerland was most frequent on host trees of sycamore maple (Acer pseudoplatanus) and beech (Fagus sylvatica); both host trees were scattered throughout their study area. Clearly, L. pulmonaria can grow in a range of situations, though its actual distribution in Tanzania is not well known and documented (Pócs Reference Pócs and Newmark1991).
Nevertheless, in many Central European regions, L. pulmonaria is often restricted to single trees due to increased anthropogenic interference at the tree and stand levels (Scheidegger et al. Reference Scheidegger, Frey, Walser, Kondratyuk and Coppins1998). In addition, existing literature indicates that L. pulmonaria is an indicator species for long ecological continuity in forest ecosystems (Gauslaa & Solhaug Reference Gauslaa and Solhaug2000). However, the presence of only three host trees of Ocotea usambarensis, which is a commercially important hardwood species in the montane forest on the southern slopes of Mt Kilimanjaro (Pócs Reference Pócs and Newmark1991; Nsolomo & Venn Reference Nsolomo and Venn2000), is explained by extensive exploitation in the past for timber (Misana Reference Misana and Newmark1991). A similar situation was observed on Mt Meru, where L. pulmonaria was found on only one Podocarpus latifolius among the few specimens of this species that were observed in the study sites, and which are commercially important hardwood species (Misana Reference Misana and Newmark1991). The scarcity of these commercially valuable tree species reflects human-induced disturbance, such as over exploitation of timber, firewood and charcoal (Giliba et al. Reference Giliba, Mafuru, Paul, Kayombo, Kashindye, Chirenje and Musamba2011), and deliberate fire setting (field personal observation).
Association of Lobaria pulmonaria with host tree variables
The results of this study suggest that host trees in sub-alpine and montane forests of Mt Kilimanjaro and Mt Meru host the largest thalli on larger, smooth bark surfaces and straight trunks. This may imply that the further colonization and occurrence of small thallus diameters on large trunks is interrupted by features associated with the wide host trees circumference such as a rapid growth of the crown, resulting in a fluctuating light (Scheidegger Reference Scheidegger1995; Gauslaa & Solhaug Reference Gauslaa and Solhaug2000; MacKenzie et al. Reference MacKenzie, MacDonald, Dubois and Campbell2001). There are other potentially important factors. Sillett et al. (Reference Sillett, McCune, Peck, Rambo and Ruchty2000) reported that small thalli were found frequently establishing on moss-free bark, and often found vigorous thalli of old-growth associated species on smooth-barked twigs and branches. Other studies have similarly reported that several microhabitat factors such as the height of the trunk, bark slope and bark moisture (Bates Reference Bates1992), and bark roughness, tree age and size (Ellis Reference Ellis2012) have an important role in determining the establishment and distribution of epiphytic lichen species in forest ecosystems (Mistry & Berardi Reference Mistry and Berardi2005; Mežaka et al. Reference Mežaka, Brūmelis and Piterāns2008; Ellis Reference Ellis2012). Also, the physical and chemical quality of the host trees’ bark structure which changes through time may enhance an old host tree in providing a different substratum than the young host tree (Lie et al. Reference Lie, Arup, Grytnes and Ohlson2009). Furthermore, it has been reported that colonization of macrolichens is often restricted to the rough branch scars as opposed to adjoining smooth bark. This is because the young bark is relatively smooth, resinous, non-absorbent and moss-free compared to the old-bark surfaces which become rough, less resinous, more porous, absorbent and moss-covered; the latter type of bark surface stores water, accumulates humus and forms a spongy surface that could either promote or inhibit lichen thallus development (Sillett et al. Reference Sillett, McCune, Peck, Rambo and Ruchty2000). In this study, we observed an enormous quantity of bryophyte mats on host trees associated with the occurrence of L. pulmonaria on both smooth and rough barked trunks. The presence of a large quantity of bryophyte mats on some host trees such as Erica excelsa, Ilex mitis, Rapenea melanophloeos and Senecio sp. in montane and sub-alpine forests suggests they might be an optimal substratum for the establishment and growth of large thalli.
Studies have found out that L. pulmonaria is restricted to bark that is not strongly buffered, with a relatively high pH (5.0–6.0) (Gauslaa Reference Gauslaa1985; Rose Reference Rose1988; Scheidegger Reference Scheidegger1995). Rose (Reference Rose1988) discovered that rain of low pH, contaminated by sulphuric and nitric acids of industrial or automobile origin, acidifies bark and may have a negative effect on thallus development. Rose (Reference Rose1988) further reported that trees with bark of higher pH (above c. 5.0) hosted L. pulmonaria, and those with low pH bark did not. This study found a high buffer capacity of bark pH between 5.1 and 7.1 and this phenomenon corresponded to some host trees such as Hagenia abyssinica having a high cover of L. pulmonaria. However, the occurrence of H. abyssinica was restricted only to Mt Meru and it was not recorded in the forests of Mt Kilimanjaro. Further investigation on the effect of bark pH, amount of moisture and mineral content as important factors for the occurrence and absence of L. pulmonaria on tree species needs to be conducted.
Conclusion and implications for lichen conservation
The nature of the present study on macrolichen species in tropical montane forests demonstrates the importance of conservation management systems that ensure forest stand continuity for lichen conservation. Based on the findings of this and related studies discussed above, it can be concluded that the current forest fire activity in protected areas may lead to the decline of regionally rare and threatened populations of L. pulmonaria. The largest trees host the largest, remnant thalli. Therefore, conservation programmes in tropical montane forests should improve fire prevention regimes and establish a monitoring programme to preserve viable populations of the target species in the protected areas.
Acknowledgements
We thank the Tanzanian Commission for Science and Technology (COSTECH), the Tanzania Wildlife Research Institute (TAWIRI) and the Kilimanjaro National Park (KINAPA), Arusha National Park (ANAPA) and Tanzania National Park (TANAPA) authorities for their great support and for granting us access to the national park areas. We would like to thank Mr Filbert Kifai Nyange, Isaya Mbumi, Gabriel Lyser (botanist) and Canisius J. Kayombo (botanist) for volunteering as field assistants and providing good company in many long hours in the forests. We wish to thank Dr Urs Kormann from Oikostat GmbH for his assistance in data analysis and interpretation. This study was conducted within the framework of research collaboration between Sokoine University of Agriculture (SUA) - Department of Forest Biology and the Swiss Federal Institute for Forest, Snow and Landscape Research (WSL) - Department of Biodiversity and Conservation Biology. Thanks to Mohamed bin Zayed Species Conservation for financial contributions during the fieldwork and WSL for financial support during two research stays in Switzerland. We also wish to thank the anonymous reviewers who greatly helped to improve the quality of the previous versions of this paper.
Author ORCID
Nuru N. Kitara, 0000-0002-5725-4621.
Supplementary Material
To view Supplementary Material for this article, please visit https://doi.org/10.1017/S0024282922000305.
