1. Introduction
As the living basis of biosphere, ecosystem productivity (EP) is easily influenced by climate changes (Ciais et al., 2003; Brookshire & Weaver, 2014; Seddon et al., 2016; Madrigal-González et al., 2017). This kind of influences has proved to become more serious along with the increases of climatic extremes (Grimm et al.,2013), particularly for the lower-EP ecosystems such as those temperature-sensitive Arctic ecosystems (Kim et al. , 2017) and the higher-EP ecosystems such as those drought-susceptible tropical rainforests (Bonal et al. , 2016). Probing the underlying rules of such ecological effects (eco-effects), especially for the lower-EP and higher-EP ecosystems with distinct responses (Grimm et al.,2013), facilitates better modeling terrestrial processes (Ruddellet al., 2016) and predicting the future of our Earth (Chen & Zhang, 2013). However, this task is quite challenging, because the ecological causes, processes, and performance of this kind of influences are so complicated – involving so diverse climatic factors and so complex interaction mechanisms (McGuire et al., 1992). This reasoning is exemplified by the case of snow phenology (SP), which involves periodic snow phenomena such as snow onset, snow end, and snow cover duration (Peng et al., 2013). People found SP may positively or negatively affect ecosystems’ performance (Lin & Hyyppä, 2019), through playing its ecological roles of adapting the chilling effect (He et al., 2018), balancing subsurface water supply (Schlaepeer et al., 2012), and keeping under-snow temperature (Zeeman et al., 2017) for organisms living. Recent SP-relevant ecological progresses reported that while in spring the timing of snowmelt partly depends on temperature, snowmelt and temperature sometimes work independently from one another to change species’ flowering timing (Steltzer et al., 2009); moreover, some species seem to be approaching their extremes of phenological shifts in response to snowmelt date rather than temperature (Iler et al., 2013). Such knowledge, however, is in shortage for ecosystems at large scales, as evidenced by the fact that the contemporarily popular dynamic global vegetation models mostly just contain simple snow functional modules (McCarthy et al., 2012; Murray, 2014; Yang et al., 2015). This has restricted the models from fully simulating climate-biosphere interactions and, eventually, from accurately projecting EP. Consequently, the related kernel macrosystem ecological (macroecological) (Heffernan et al., 2014) questions such as “how may SP influence EP, particularly for those lower- and higher-EP ecosystems?” have been extensively asked, and seeking their answers has attracted a lot of attention in the fields of ecology, global carbon cycles, and Earth interactions (Steltzer et al., 2009; Schlaepeeret al., 2012; Peng et al., 2013).
For the question, we further burst with a curiosity on its commonly-assumed premise – “must this eco-effect be caused by the SP in the same site exclusively?” or “may EP be more influenced by the SP around?”. So far, no answers have been found, let alone for the special cases of lower- and higher-EP ecosystems. The raising of such an odd question stemmed from our re-examination of the first question by analogy with the principle-approximate problems in the temporal field. For example, for understanding how temperature controls leaf onsets (Wuet al., 2015) or snow cover accumulations (Lin & Jiang, 2017), scientists were often devoted to deriving the temporal lengths of “preseasons”, whose definition is the phase with the highest correlation between the two feature series of a periodic event of interest and its driver at the different phases preceding the occurrences of that event (Lin & Jiang, 2017). The “preseason” here, in effect, is analogous to the “distance” of the neighboring site that acts with the greatest eco-effect. In such a sense, it is reasonable to pose the second question. Actually, this analogy had its theoretical foundation, i.e., the spatial gradients (this term already involving the adjacent sites) of temperature tend to affect the dynamics of snow accumulations (Yang et al., 2011), mechanically through atmospheric circulations (Li et al., 2013; Ohara et al.,2014; Johnson & Ohara, 2018). Thereby, the scheme and strategy of characterizing the time-lag eco-effect (Wu et al., 2015; Lin & Jiang, 2017) seemingly can be transferred into the spatial field for detecting the possible eco-effects from the neighboring sites. This analogy is why the second question was posed. Altogether, the two questions can be schematically shown in Fig. 1, and their solving can be methodologically specified as a task of exploring the strongest correlation between the interannual series of EP values at each site and the related interannual series of its corresponding/surrounding site-related snow-onset/end SP values. Completing this task, no doubt, can enrich the ecological knowledge of SP-EP interactions, enhance snow functional modules in dynamic global vegetation models (McCarthyet al., 2012; Murray, 2014; Yang et al., 2015), and, further, improve the accuracies of climate-based EP predictions (Chenet al., 2013).
-Insert Fig. 1 here-
To answer the two questions, we innovatively proposed a new concept of “periconnection”, by referring to the way of defining “teleconnection” (Mikolajewicz et al., 1997) but also expandingly exploring the “connections” to neighborhoods, for reflecting the potential links between EP and its corresponding and neighboring site-related SP. Teleconnection refers to climate anomalies being related to each other at large distances (typically for thousands of kilometers) (Allen & Luptowitz, 2017). The general initiative of teleconnection studies is like first “hunting” (exploratory analysis) statistically significant correlations, and then, varying methods of correlation analysis are used to determine the existence and degrees of teleconnections (Mamalakis et al., 2018). In fact, such a conceptual reference to “teleconnection” has emerged in the field of urban ecology (Seto et al., 2012), but no defined concepts in this direction have been reported. We attempted to fill this gap, thru a case study of exploring the macroecological links between the Arctic circumpolar EP and the corresponding/surrounding site-related SP in the Northern Hemisphere (>45°N) (NH).