Implications for diversity gradients in general
Looking at different diversity gradients and generalising the findings might be a chance to get a deeper insight into mechanisms of species richness. The depth diversity gradient (DDG) of macrophytes brings some advantages compared the latitudinal diversity gradient (LDG) or the elevational diversity gradient (EDG). The DDG is formed for shorter scales (few meters) than the EDG, which implies a lower importance of dispersal or connectivity processes. Another big advantage looking at freshwater depth pattern is that it is much easier to use replicates for DDG than for LDG and EDG. For LDG, the options for replicates are restricted as there are mainly two replicates for the two hemispheres, which can be further (pseudo)replicated when studying different taxa or continents (Pontarp et al. 2019). For EDG, comparative studies require a high logistic sampling effort. Thus, studies addressing the overall shape across EDGs and the variables affecting them are not yet clear (Nogués-Bravo et al. 2008, McCain and Grytnes 2010, Kessler et al. 2011, Sanders and Rahbek 2012). Hence, the smaller scales of DDG seem an advantage to do such replicate studies to address the role of energy or productivity in shaping the diversity gradients. Moreover, a comparative study across different gradients (EDG vs. DDG) could yield further insights in understanding diversity drivers.
Besides DDG, there are other small-scale diversity gradients that get studied. One example is the height or vertical diversity gradient (VDG) from forest floor to tree crowns following environmental gradients of light intensity, temperature and humidity. Here, species richness distribution depends on the study species group. Hump-shaped patterns for ferns and lycophytes (Acebey et al. 2017) or vascular epiphytes (Krömer et al. 2007) or decreasing patterns for vascular epiphytes (Wang et al. 2016) were found. The pattern does not seem to be very clear and are strongly influenced by host crown characteristics (Wang et al. 2016). Still, VDG could be the most comparable gradient to DDG for contrasting the role of light and temperature, but paired data with both diversity and environmental data for VDG is even scarcer than for DDG.
The LDG for species generally peak in the tropics (Pontarp et al. 2019). However, looking at freshwater plants, the diversity gradient does not decrease linearly (Alahuhta et al. 2020), but seems to have at sub-tropical to low tropical latitudes (Murphy et al. 2019). Thus, along latitude, the species richness peak is at a high to intermediate level of solar productivity, which is reflected in our analysis of DDG. This might be explained by the Intermediate Productivity Hypothesis (IPH). The IPH, stating that at low productivity level (deep waters with low light quantity and low temperature) only few species can survive and at high productivity level (shallow waters with high light quantity and high temperature) only few competitive species survive, is highly discussed (Adler et al. 2011, Fraser et al. 2015) and mostly used along LDG. Besides competition, mid-gradient peaks can be explained from an evolutionary perspective as a maximal difference between speciation and extinction rates is expected at intermediate productivity (VanderMeulen et al. 2001). Although quantification of productivity along depth should be attempted, our findings already indicate a key role of light quantity and temperature in shaping DDG.
The hump-shaped curve of the DDG might mirror most elevational diversity gradients (EDG), which often peak at intermediate levels (Kessler et al. 2001, Nogués-Bravo et al. 2008). For both gradients one explanation might be geometry. The mid-domain effect (MDE) explains richness peaks in the middle of gradients by species range overlaps simply via geometric constraints (Colwell et al. 2004). This concept is mainly used to explain the EDG (McCain 2004, Cardelús et al. 2006, Brehm et al. 2007). Also along the depth gradient niches of single species might overlap and create an area of enhanced richness. Nevertheless, to adequately verify the influence of the mid-domain effect, it would be necessary to study single species in relation to environmental variables via mesocosm experiments without competition. Potential niche ranges associated with lake depth (e.g. light or temperature conditions) could then be determined and the overlap of these ranges for all co-occurring species of a particular lake would provide at which depth lies the MDE-predicted maximum number of species. This would make it possible to compare observed vs. MDE-predicted DDGs. Moreover, another important explanatory theory of EDG is the Intermediate Disturbance Hypothesis (IDH). It suggests that species richness is highest at mid-levels of disturbance as species of early and late successional phases can coexist (Connell 1978). Whereas disturbances along EDG are caused by enhanced human activities at lower elevation (Nogués-Bravo et al. 2008), disturbances along DDG are caused by waves, grazing and water level fluctuations at shallower depth. These disturbances act on two different depth patterns: 1) evenly distributed over depth like herbivory by fish or 2) gradually distributed (loosing strength with depth), which can be caused by anthropogenic use, herbivory by birds or mammals, wind, waves (Van Zuidam and Peeters 2015), mechanical disturbance by ice cover, or water level fluctuations (Evtimova and Donohue 2016). The latter was already integrated in our study in a very simple way of taking long term indicators. Further monitoring schemes should target other disturbances, quantifying herbivory, wave height, ice cover or finer resolved water level fluctuations. As herbivores are supposed to eliminate on average 40-48% of plant biomass in freshwaters (Bakker et al. 2016) their influence might be outstanding. The long-term water level fluctuation showed no strong effect on richness measures and it might not be the main explanation of the hump-shaped DDG. Nevertheless, although our findings tend to not support the intermediate productivity/disturbance hypotheses, considering that human (or natural) disturbances in shallow waters should be more prevalent in our data, the intermediate productivity/disturbance hypotheses could still play a role along DDG.