Epiphytic bryophytes are some of the key species characterizing mid-altitude tropical montane cloud forests (Bruijnzeel et al., 2011b) and play a pivotal role in influencing the global hydrological cycle (Porada et al., 2018). Due to the morphology of the species, it is very difficult to quantify the abundance. In this study, we demonstrate that it may be possible for regional mapping EB biomass. However, challenges and uncertainties remain and need to be assessed. Our discussion will mainly focus on (1) EB depth-biomass allometry; (2) scaling of EB biomass from the patch to forest stand scales; (3) determinants governing the abundance of EB and (4) remotely sensed regional EB biomass estimation.
doi:10.6342/NTU201902408
4.1 Epiphytic bryophytes depth-biomass allometry
Plant allometry focuses on relationships between plant body size and biomass, production, population density or other dependent variables (Enquist et al., 1998; Enquist et al., 1999).
Stanton and Reeb (2016) pointed out that some characteristics of bryophytes may be allometrically scaled like vascular plants. However, it may not be applicable to all species.
In this study, in-situ general allometric equations were developed to estimate EB biomass using the central depth of sampled 100 cm2 circular patch. The mean exponent of the power models of these five equations was 0.75 (3/4) ranging from 0.72 to 0.83 (Table 2), which agrees with the 3/4 power law (Kleiber, 1947) and is similar to the constant scaling exponents over a wide range of vascular plant size often with a quarter-powers in metabolic scaling theory (West et al., 1997, 1999). However, epiphytic bryophytes are non-vascular plants which compose of a simple stem has only a limited role in transporting moisture and nutrients through conducting tissues and they do not follow the vascular transport system as a self-similar, fractal-like branching network (Ligrone et al., 2000). Two major branching forms in bryophytes are sympodial with connected modules of same level and monopodial (Stanton & Reeb, 2016). For most vascular plants, the branching bifurcation is two (Enquist et al., 2007), and the height is 1/4 exponent of mass (West et al., 1999), which was different to our empirical observation. This verifies that the basic assumption of self-similar branching network of an organism plays a major role in governing the allometric relationship.
31
4.2 Scaling of EB biomass from the patch to forest stand scales
It is extremely challenging to non-destructively measure EB biomass, and a new field approach was invented in this study (Figure 3). This is crucial since it could take many years for EB to recover (Fenton et al., 2003). In addition, the approach was efficiently, which may permit fast sampling for a large region. This study focuses on the height below 3 m where the majority of EB are present (Trynoski & Glime, 1982). The sampled area may be further extended with aids of foldable ladder or tree climbing. With the availability of previously derived in-situ ratio of biomass of the sampled area and the total area of a tree, we may be able to estimate EB biomass at the individual tree and the plot scales. This sampling approach should be reliable and the models explained 72% of data variation (Figure 4). The performances were satisfactory since only a single variable (the central depth of a sample) was used to model biomass of EB with variable morphologies.
In this study, assumptions were made that the sample circular areal size of 100 cm2 was appropriate and the central depth was representative, and these should be further investigated especially for the latter one. In some cases, plant depths within the circular shaped frame were not uniform mostly due to the variation in the abundance. In some rare cases, more than one species were present within the sampling, which were not visible on site. Hence, the use of central depth could be dubious and may induce model uncertainty. At the tree scale, the distribution of EB biomass on a tree was influenced by microclimate positions and directions. Therefore, we calculated the depth of EB in four and eight directions for small and large trees, respectively, which may reduce microclimate-induced biases (Figure S4). As we extrapolated EB biomass of an entire tree, an in-situ unfixed coefficient was applied to convert EB biomass of the sampled stem area to EB biomass of an entire tree. One potential research limit is that the tree
doi:10.6342/NTU201902408
scale EB biomass estimation, which was extrapolated from it of the sampled tem area and was interpolated to the plot scale, was not validated. This requires tree climbing or destructive harvesting to clear EB of entire trees. Unfortunately, there was no availability of the support of tree climbing. Logging in nature forests has been completely forbidden in Taiwan since 1991. Therefore, the latter option may not be possible due to the local regulation. In the future, we might be able to take the advantage of tropical cyclone-induced fallen logs and harvest EB biomass at the ground level.
33
4.3 Determinants governing the abundance of EB
In this study, we compiled a large set of biophysical, topographic and bioclimatic factors derived from field and airborne lidar data to investigate salient factors that govern the abundance of EB in TMCF of Chilan Mountain in northern Taiwan. For biophysical determinants, as expected, the parameters that were related to tree size (aboveground biomass, DBH and CHM) were positively related to TBB. The larger DBH of host tree provide more diverse micro-habitats for EB succession resulting from different bark structure, trunk chemical composition, humidity and luminosity (Chen et al., 2010). We note that data of a small portion (2 of the 21 [11%]) of the field plots (plot FR170 and Neighbor) with high biomass density (≥330 Mg ha-1) deviated from the main pattern, and negative trends of TBB and aboveground biomass, DBH and CHM were observed (Figure S5). This suggests that there may be a threshold for the positive relationship between TBB and tree size at the upper end. Sites with extremely high carbon storage of this study region may be characterized as relatively lower and the higher luminosity resulting in a desiccative micro-habitat and harbored fewer TBB.
For topographic attributes, TBB generally increased with elevation in 50% of the surveyed plots. The finding agreed with Wolf (Wolf, 1993) with the estimation of bryophyte biomass across an elevation gradient in Andes. However, intuitively, lower elevations of the TMCFs within the cloud band should experience more daily upslope fog from moist wind from the Pacific Ocean, and may be able to harvest sufficient cloud water in the study region. The complex relationships among slope, profile curvature, plan curvature and TBB were difficult to decipher (Werner et al., 2012). The results indicated that the slope, profile curvature and plan curvature needed to be taken into account, which
doi:10.6342/NTU201902408
may be related to soil water contents, solar radiation, flow acceleration and erosion (Moore et al., 1991). The southwestness was also related to the intensity of solar radiation and the high frequency of the upslope fog (Błaś et al., 2002; Wang et al., 2016). Overall, southeast-ward wind during the day affected by terrain resulting in much frequent occurrence of upslope fog at southeast forest stands bring caused much moist microclimatic environment in Chilan Mountain. However, the southwestness was not a salient factor related to TBB. We inferred that the slight difference of southwestness within plots had little effect on TBB owing to the abundant rainfall in this region. In addition, the difference of southwestern may not be the main influence of luminosity due to the similar canopy closure in most plots except for the extreme DBH large stands.
For bioclimatic attributes, air temperature was a salient factor and negatively related to TBB. In fact, the results reflected that the elevation was linearly and negatively correlated with temperature. The abundant precipitation (rainfall and cloud water) (4000+
mm annually) of the study region may play a minor role in regulating the abundance of EB and water related bioclimatic attributes may be overlook. In the future, under-canopy solar radiation and fog occurrence (from a time-lapse video) data may be included for a more comprehensive analysis.
35
4.4 Remotely sensed regional EB biomass estimation
Regional estimation of EB biomass in TMCF is challenging and the spatial variation in abundance may be governed by biophysical, topographic and bioclimatic factors (Wolf, 1993). By investigating these salient factors, we developed a spatial explicit model to map regional EB biomass of a TMCF. Our regional EB biomass estimates of 188 100 (mean standard deviation kg ha-1) ranging from 0 to 650 kg ha-1 (Figure 8) were similar to previously reported values (Table 4). We note that only the EB biomass was estimated in this study, and the bryophytes on the forest floor were not taken into account. The same field and airborne lidar approaches should be feasible for the estimation, and bryophytes of the entire regional could be assessed. Although airborne lidar has been heavily utilized for ecological research for more than two decades (Lefsky et al., 1999), high cost is still the main constraint for the data acquisition. With the potential of satellite high-resolution laser ranging from the Global Ecosystem Dynamics Survey (GEDI) that is available for the public (Dubayah et al., 2014), the proposed approach may become even more valuable.
doi:10.6342/NTU201902408
Table 4 Summary of relevant bryophytes biomass research.
Study Location Vegetation type Elevation (m a.s.l.)
37