The Role of Cyanobacterial External Layers in Mass Transfer: Evidence from Temperature Shock Experiments by Noninvasive Microtest Technology

Groundwork on cyanobacterial external layers is crucial for an improved understanding of the persistent dominance of cyanobacteria in freshwaters. In this study, the role of two morphotypes of external layers in Microcystis and Nostoc in mass transfer and instantaneous temperature shock were explored by noninvasive microtest technology (NMT) after a series of pretreatments, to obtain the external layers retained or stripped samples. The results showed no statistical influence on photosynthetic activity between retained and stripped samples in both Microcystis and Nostoc. External-layer-retaining strains had higher net O2 effluxes than stripped strains. Moreover, the net NH4+ influx was significantly higher for the sheath retaining Nostoc than for the stripped sample, indicating that external layers might be an important feature driving mass transfer in cyanobacteria. However, the role of slime in NH4+ absorption was limited compared with that of sheath. In addition, external-layer-retaining strains exhibited a longer response time to instantaneous temperature shock, greater net O2 effluxes at a 4 °C shock and lower net O2 influx at a 35 °C shock, which were interpreted as reflecting a tolerance to temperature fluctuation over short time scales via a buffer function of external layers to stabilize cell activity, ameliorating the efficiency of photosynthesis and respiration. These results advance current knowledge regarding the external layers, especially the dense sheath, involved in the mass transfer in cyanobacteria, and provide new clues concerning the adaptive strategies of cyanobacteria under global climate changes.


Introduction
Surface coats are widespread, surrounding microorganisms as well as some plant and animal cells. By comparing the surface coats of a variety of cell types, Bennett [1] proposed a generalized terminology of "glycocalyx" for this biochemical structure. Over the past decades, this structure has received increasing attention; yet, the terminology is often confused and not strictly followed. Initially, it was referred to as the sheath and/or capsule based on morphological aspects in microorganisms [2,3]. The terms mucilage and slime were also used in some earlier publications [4,5]. Later, because the microbial surface layer is mainly composed of carbohydrates and can be secreted outside the cell, it was termed the capsular polysaccharides (CPS) [6]. In recent years, some researchers have further divided this structure into loosely bound exopolysaccharides (LB-EPS) and tightly bound exopolysaccharides (TB-EPS) in terms of binding to cells [7,8]. Instead, regardless of the differences in the definitions above,

Strains and Culturing
Two Microcystis species (M. aeruginosa FACHB-1338 and M. sp. FACHB-2427) and two Nostoc species (N. sp. FACHB-599 and FACHB-2009) were used in this study (Table 1). They were obtained from the Freshwater Algae Culture Collection of the Institute of Hydrobiology, Chinese Academy of Sciences (FACHB-Collection, Wuhan, China). These strains were all kept in the colony phenotype, coating with different types of external layers. The experiments were conducted in BG11 medium under a cool-white fluorescent light intensity of 50 µmol m −2 s −1 with a 12-h: 12-h light: dark cycle at a temperature of 25 ± 1 • C. The culture conditions remained the same throughout the study.

Experimental Design
The first phase of the study was to obtain the external-layer-retaining and stripped samples of four species. Colonial M. aeruginosa FACHB-1338, M. sp. FACHB-2427, N. sp. FACHB-599 and FACHB-2009 were grown in triplicate in 50-mL Erlenmeyer flasks separately. All cultures were manually shaken 3 or 4 times every day, and their position was randomized daily. Cells in the exponential growth phase in batch mode were harvested by centrifugation, and then resuspended and divided into equal halves, one for colonial samples with external layers retained, and the other for the stripped samples. To retain cell activity, Microcystis and Nostoc colonies were first centrifuged at 5000 g for 10 min, and suspended in a 0.05% NaCl solution to the original volume [7]. Low-dose ultrasound was used in our study to avoid overexposure and to achieve disaggregation of colonies and removal of the external layers [28]. Nostoc strains were treated by ice bath ultrasonication for 60 s at 20 kHz and 80 W (CPX-130, USA) to obtain stripped samples. Microcystis strains were subjected to ultrasound at 30 W for 60 s and then centrifuged at 12,000× g for 20 min (4 • C) to obtain stripped samples. Then, these external-layer-stripped cells were observed with India ink staining by light microscopy to check the sheath and/or slime.
In the second phase, NH 4 + and O 2 fluxes of external-layer-retaining and stripped samples of Nostoc and Microcystis were determined by NMT, to evaluate any changes after the removal of the sheath and/or slime. The external-layer-retaining and stripped samples of Nostoc and Microcystis were resuspended in BG11 N-medium, at a cell density of approximately 10 7 cells mL −1 , followed by culturing under routine conditions for 24 h. The photosynthetic activity of these strains was measured, and, subsequently, 1 mL culture samples were placed in the middle of poly-L-lysine-pretreated coverslips (2 × 2 cm), in a measuring chamber consisting of a glass petri dish (35 mm). After the cells had settled on the coverslips, 5 mL NMT measuring solution was gradually added to the glass petri dish. NH 4 + and O 2 fluxes of all Nostoc and Microcystis samples were then monitored by NMT, respectively.
In the third phase, to further understand the role of the sheath and/or slime in stress tolerance, an instantaneous temperature shock experiment was designed, in which M. sp. FACHB-2427 and N. sp. FACHB-2009 including two phenotypes, were used. The procedure involved the following: (i) the measuring chamber mentioned above was fixed in the center of another 100-mm glass petri dish; (ii) O 2 fluxes were monitored at room temperature until the curve was stable; and (iii) water at different temperatures was rapidly added into the 100-mm glass petri dish separately, and an ice bath was applied at 4 • C, and a hot bath at 35 • C in the measuring chamber, respectively, with continuous O 2 flux monitoring. All the NMT experiments described in this section were determined by measuring at least six similar samples separately, and each measurement was repeated three times at different positions of cell.

Light Microscopy
External-layer-retaining and stripped samples of Nostoc and Microcystis were observed with an inverted Olympus microscope (Olympus IX73, Japan) before and after staining with India ink.

Photosynthetic Activity Determination
In vivo chlorophyll fluorescence was measured with a phytoplankton analyzer (PHYTO-PAM, Walz GmbH, Germany). All strains were dark-adapted for at least 15 min, before measuring the fluorescence parameters (photosystem II activity, PSII). The maximum effective quantum yield of PSII was calculated according to the following equation: [29], where F m and F 0 represent the maximum and minimum fluorescence values of the dark-adapted stage of PSII, and F v is the difference between them. 0.5 mM) to choose the qualified one with a Nernstian slope at 58 ± 5 mv/decade. The microsensor was then placed in the blank measuring solution for testing by X-10 until the NH 4 + flux was near the baseline, at which time the microsensor can be used [30].
To detect dissolved oxygen, the Pt/Ir polarographic oxygen microsensor (tip diameter 20 ± 5 µm, XY-CGQ-501, Younger USA) was used under -750-mV polarization voltage. A reference microsensor was also used to complete the circuit. Prior to O 2 flux measurement, the microsensor should be calibrated with measuring solution (0.1 mM KCl, 0.1 mM CaCl 2 , 0.1 mM MgCl 2 , 0.5 mM NaCl, 0.3 mM MES, 0.2 mM Na 2 SO 4 , pH 6) containing different concentrations of O 2 (N-saturated and control cultural media). Only when the Std Curve is between −2000~−9000 pA/mM can the microsensor be placed in the blank measuring solution for polarization for 1 h, until the net O 2 flux is near baseline, at which time the microsensor can be used [31].
During formal measurement, the fluxes of NH 4 + and O 2 were determined by measuring six similar samples separately. The potential difference was obtained by moving the microelectrode repeatedly from one point to another, in a direction perpendicular to the surface of the individual cell (Polar X-10), and fluxes were calculated automatically by Fick's law of diffusion: J = −D(dc/dx). The steady-state flux measurements were continuously recorded for 6-10 min, and each measurement was repeated three times at different positions of the cell.

Data Processing and Statistical Analysis
The data obtained from NMT were exported as raw data and then converted into net fluxes by JCal V3.3 (a free MS Excel spreadsheet, http://www.youngerusa.com). For analyses of net NH 4 + and O 2 fluxes of external-layer-retaining and stripped samples of Nostoc and Microcystis, readings were averaged to obtain the net ionic and molecular steady fluxes for 6 min at each measurement position in each sample. The coefficient of variation (CV) was calculated as the standard deviation divided by the mean value. To determine the response time to instantaneous temperature shock, response curves of net O 2 fluxes over 10 min were the first 5 points adjacent-averaging smoothed [32], and steady fluxes for 100-200 s, as appropriate, before temperature shock were averaged and regarded as the initial level. After instantaneous temperature shock, gradient recovery within ±3 pA was considered as stable data. When the smooth curve reached the initial level again, the relevant time was considered the end time of the response. Readings within the response time were integrated, and the value of integration versus time was adopted as the weighted average flux, in response to instantaneous temperature shock. Data in this study are presented as the mean ± standard error (SE). The results of the experiment were analyzed by ANOVA, using Tukey's post hoc test. All statistical analyses were carried out with Origin 9.0 (OriginLab, USA). Differences were considered significant at p < 0.05.

Evaluation of External Layer Extraction
Optical microscopy observations showed that two strains of Nostoc were characterized by the presence of a firm sheath surrounding the filaments. Within the sheath, the filaments were "randomly" loosely arranged, and they showed an irregularity of coiling in N. sp. FACHB-599, whereas filaments in N. sp. FACHB-2009 were more tightly packed (Figure 1a,b). A mucilaginous layer of slime outside the colony was observed by India ink staining in two Microcystis strains. Multiple unicells were loosely assembled in the colony and remained irregular in shape (Figure 1c,d). After removal of the sheath and slime, the morphologies of Nostoc and Microcystis were short filaments containing 4-30 cells and single-cell forms, respectively. In these external-layer-stripped samples, neither the sheath nor the surrounding slime was observed by India ink staining, which suggests that the external layer removal was effective (Figure 1e-h).

Data Processing and Statistical Analysis
The data obtained from NMT were exported as raw data and then converted into net fluxes by JCal V3.3 (a free MS Excel spreadsheet, http://www.youngerusa.com). For analyses of net NH4 + and O2 fluxes of external-layer-retaining and stripped samples of Nostoc and Microcystis, readings were averaged to obtain the net ionic and molecular steady fluxes for 6 min at each measurement position in each sample. The coefficient of variation (CV) was calculated as the standard deviation divided by the mean value. To determine the response time to instantaneous temperature shock, response curves of net O2 fluxes over 10 min were the first 5 points adjacent-averaging smoothed [32], and steady fluxes for 100-200 s, as appropriate, before temperature shock were averaged and regarded as the initial level. After instantaneous temperature shock, gradient recovery within ± 3 pA was considered as stable data. When the smooth curve reached the initial level again, the relevant time was considered the end time of the response. Readings within the response time were integrated, and the value of integration versus time was adopted as the weighted average flux, in response to instantaneous temperature shock.
Data in this study are presented as the mean ± standard error (SE). The results of the experiment were analyzed by ANOVA, using Tukey's post hoc test. All statistical analyses were carried out with Origin 9.0 (OriginLab, USA). Differences were considered significant at p < 0.05.

Evaluation of External Layer Extraction
Optical microscopy observations showed that two strains of Nostoc were characterized by the presence of a firm sheath surrounding the filaments. Within the sheath, the filaments were "randomly" loosely arranged, and they showed an irregularity of coiling in N. sp. FACHB-599, whereas filaments in N. sp. FACHB-2009 were more tightly packed (Figure 1a,b). A mucilaginous layer of slime outside the colony was observed by India ink staining in two Microcystis strains. Multiple unicells were loosely assembled in the colony and remained irregular in shape (Figure 1c,d). After removal of the sheath and slime, the morphologies of Nostoc and Microcystis were short filaments containing 4-30 cells and single-cell forms, respectively. In these external-layer-stripped samples, neither the sheath nor the surrounding slime was observed by India ink staining, which suggests that the external layer removal was effective (Figure 1e-h). The extraction methods mainly consisted of physical processes, including ultrasound and centrifugation. An excessive intensity and duration of ultrasound is known to cause rapid and severe cell disruption and photosynthesis inhibition [33]. In general, F v /F m is used as a sensitive indicator of photosynthetic performance, in response to environmental stress [34]. The different F v /F m values between species reflect their distinct differences in the potential quantum efficiency of PSII. Changes in F v /F m , resulting from ultrasound assisted extraction, are shown in Figure 2. In N. sp. FACHB-599 and FACHB-2009, the ratio of F v /F m deceased slightly in the sheathless strains, compared with the colonial ones, while the F v /F m value increased in the slimeless strains of M. aeruginosa FACHB-1338 and M. sp. FACHB-2427. However, no significant difference was observed in the F v /F m value between all the external-layer-retaining and stripped strains (ANOVA, p > 0.05), indicating that the low-frequency and power of the ultrasound applied in our study did not induce physiological deactivation of Nostoc and Microcystis. This result is consistent with the findings of Francko et al. [35], who concluded that low-dose ultrasound (50 W, 20 kHz) could provide an environmentally safe method for enhancing cyanobacterial growth. In contrast, Zhang et al. [36] reported that sonication effectively damaged cyanobacterial photosynthesis. We speculate that these discrepancies might be due to the variation in the ultrasonic conditions employed, and different morphologies of species. Purcell et al. [37] proposed that the susceptibility of microalgae to ultrasound may vary, depending on the morphological differences in shape and cell wall structure. In our experiments, the presence of different external layers was also a probable reason for various physiological changes. Nevertheless, the above results supported our assumptions that these ultrasonic treatments did not cause a significant eco-physiological change in our cyanobacteria samples, providing the basis for follow-up analysis and discussions. The extraction methods mainly consisted of physical processes, including ultrasound and centrifugation. An excessive intensity and duration of ultrasound is known to cause rapid and severe cell disruption and photosynthesis inhibition [33]. In general, Fv/Fm is used as a sensitive indicator of photosynthetic performance, in response to environmental stress [34]. The different Fv/Fm values between species reflect their distinct differences in the potential quantum efficiency of PSII. Changes in Fv/Fm, resulting from ultrasound assisted extraction, are shown in Figure 2. In N. sp. FACHB-599 and FACHB-2009, the ratio of Fv/Fm deceased slightly in the sheathless strains, compared with the colonial ones, while the Fv/Fm value increased in the slimeless strains of M. aeruginosa FACHB-1338 and M. sp. FACHB-2427. However, no significant difference was observed in the Fv/Fm value between all the external-layer-retaining and stripped strains (ANOVA, p > 0.05), indicating that the lowfrequency and power of the ultrasound applied in our study did not induce physiological deactivation of Nostoc and Microcystis. This result is consistent with the findings of Francko et al. [35], who concluded that low-dose ultrasound (50 W, 20 kHz) could provide an environmentally safe method for enhancing cyanobacterial growth. In contrast, Zhang et al. [36] reported that sonication effectively damaged cyanobacterial photosynthesis. We speculate that these discrepancies might be due to the variation in the ultrasonic conditions employed, and different morphologies of species. Purcell et al. [37] proposed that the susceptibility of microalgae to ultrasound may vary, depending on the morphological differences in shape and cell wall structure. In our experiments, the presence of different external layers was also a probable reason for various physiological changes. Nevertheless, the above results supported our assumptions that these ultrasonic treatments did not cause a significant eco-physiological change in our cyanobacteria samples, providing the basis for follow-up analysis and discussions.

Comparison of Net NH 4 + and O 2 Fluxes of External-Layer-Retaining and Stripped Samples
To understand whether external layers influence the nitrogen uptake and photosynthetic oxygen evolution of Nostoc and Microcystis, the NMT technique was employed to monitor net unit chlorophyll compared with the unicellular form. Caution must be observed in making inferences from our results, because of the changes in size and morphology of cyanobacteria after external layer extraction. It has been generally accepted that photosynthetic parameters and nutrient uptake are related to the phenotype of cyanobacteria [38,39]. However, in the present study, the NMT platform only allowed the positioning of microelectrodes at a point near a single cell, regardless of the phenotype, e.g., colonial, filamentous, or unicellular forms. Under such an experimental microenvironment, differences in net NH 4 + and O 2 fluxes among the phenotypes with varying size were not significant (Table S1). The readily visible difference between unicellular/filamentous and colonial forms in our NMT microenvironment was the absence in external layers of the former. Thus, mass transfer of retaining and stripped samples of cyanobacteria at the submarginal level were comparable in our NMT experiments. We believe that the existence of external layers is an important feature driving O 2 efflux in cyanobacteria.

The responses of External-Layer-Retaining and Stripped Samples to Instantaneous Temperature Shock
Cyanobacterial external layers, which are mainly composed of complex heteropolysaccharides, play an important physiological role in bloom formation and various types of stress tolerance during adverse conditions [9]. In this study, the real-time net O2 flux in N. sp. FACHB-2009 and M. sp. FACHB-2427 in external-layer-retaining and stripped samples were detected under instantaneous temperature shock at 4 °C and 35 °C, respectively (Figure 4a,b). The results showed that both retaining and stripped samples of Nostoc and Microcystis displayed remarkable O2 effluxes at 4 °C shock and O2 influxes at 35 °C, indicating a significant stimulation of respiration by 35 °C compared with 4 °C shock. We infer that this difference is due to the tendency of respiration to be more sensitive  (Figure 3e). These findings, again, supported the important role of external layers in cyanobacteria in ion and molecule fluxes, whereas such effects may vary, depending on the morphotype of the external layers. One of the functions of external layers in cyanobacteria may facilitate their absorption of essential nutrients that are present in the surrounding medium at submarginal levels [21]. Unlike Nostoc colonies with a sheath form covered by a fibrillar structure, the colonial Microcystis exhibits a thick mucilaginous matrix, displaying a diffuse and loosely bound structure [12], leading to the apparent differences in the diffusion boundary layer associated with nutrient transport at the cell surface [40]. The poor contribution of slime to NH 4 + influx in Microcystis in our study was probably due to the decreased diffusive conductance of the boundary layer around colonies, compared with isolated slimeless unicells [41]. Thus, the effective diffusivity and storage of NH 4 + may not be a significant feature of the slime surrounding Microcystis in comparison to Nostoc.

The Responses of External-Layer-Retaining and Stripped Samples to Instantaneous Temperature Shock
Cyanobacterial external layers, which are mainly composed of complex heteropolysaccharides, play an important physiological role in bloom formation and various types of stress tolerance during adverse conditions [9]. In this study, the real-time net O 2 flux in N. sp. FACHB-2009 and M. sp. FACHB-2427 in external-layer-retaining and stripped samples were detected under instantaneous temperature shock at 4 • C and 35 • C, respectively (Figure 4a,b). The results showed that both retaining and stripped samples of Nostoc and Microcystis displayed remarkable O 2 effluxes at 4 • C shock and O 2 influxes at 35 • C, indicating a significant stimulation of respiration by 35 • C compared with 4 • C shock. We infer that this difference is due to the tendency of respiration to be more sensitive to temperature and to increase more than photosynthesis [42]. We then smoothed the response curves of net O 2 fluxes over 10 min using the five point adjacent-averaging method. The temperature responses of the net O 2 flux in all strains were unimodal, with rates rising up to a peak and declining thereafter at 4 • C shock (Figure 4c,d), whereas the complete reverse trend was observed at 35 • C shock (Figure 4e,f). Eventually, they all gradually returned to the initial state, which suggests that Nostoc and Microcystis have a temperature fluctuation tolerance over short time scales.
Based on the smoothed response curves of net O 2 flux, the response time to instantaneous temperature shock of external-layer-retaining strains was significantly longer than the stripped strains at both 4 • C and 35 • C (ANOVA, p < 0.05) (Figure 4c Table 2. The retaining strains of N. sp. FACHB-2009 and M. sp. FACHB-2427 showed greater net O 2 effluxes than the stripped strains at 4 • C shock (ANOVA, p < 0.05), probably because retaining strains can have a higher photosynthetic rate during cold adaptation. Compared with the sheathless strain of N. sp. FACHB-2009, the net O 2 influx in the sheath strain was significantly lower at 35 • C shock. However, for M. sp. FACHB-2427 at 35 • C shock, the net O 2 influx in the slime strain was slightly lower than the slimeless strain, although this difference was not significant (ANOVA, p > 0.05). Crucially, the respiration rate of phytoplankton rises more rapidly with increased temperature than the photosynthetic rate, resulting in universal declines in the rate of carbon fixation with short-term increases in temperature [43,44]. Through a comparative analysis of experimental data, it could be considered that cyanobacteria embedded in external layers have the advantage in thermal adaptation via downregulation of the respiration rate, thereby increasing the potential for carbon allocation to growth [42]. The external layers are fundamental to this adaptability especially in soil crust cyanobacteria. Previous work has indicated that some crust-forming cyanobacteria increased EPS secretion when subjected to diurnal temperature cycles [45]. This could also help explain why desert cyanobacteria, such as EPS-rich Nostoc colonies, can grow well when undergoing large temperature fluctuations on a daily basis [46]. Our findings suggest that external layers, especially the dense sheath, may therefore have an ameliorating impact on the efficiency of photosynthesis and photosynthesis-coupled respiration in cyanobacteria when suffering short-term temperature fluctuation.

Contribution of External Layers to the Dominance of Cyanobacteria
Increasing concern about cyanobacterial blooms worldwide has motivated research on their external layers and EPS secretion [9]. Cyanobacteria possess a variety of competitive advantages against their opponents, which allows them to be persistently dominant in freshwaters, most of which are considered to be related to the external gel-like matrix surrounding the cells. For instance, a coating of extracellular polysaccharidic material is involved in regulating buoyancy, chelating necessary metal cations, blocking chemical contaminants, and resisting turbulence [14,47,48]. Many studies, in fact, have shown that cyanobacterial external layers or bound EPS are the main contributors to colony formation, which would influence the development of cyanobacterial blooms [49]. Xu et al. [7] also found that stripped and removed bound EPS decreased cohesion and aggregation abilities by changing the surface properties of cyanobacterial cells, leading to the destabilization of cyanobacteria in water. Therefore, understanding the behavioral characteristics and functions of cyanobacterial external layers could be key to preventing bloom formation.
Typically, the presence or absence of an external gelatinous layer is considered an important adaptation mechanism of cyanobacteria to their environment [50]. As previously reported, the addition of minute increments of inorganic nutrients may have substantial effects on the growth of cyanobacteria in natural waters, but it has no effect in laboratory media, because the gelatinous layer is usually quickly lost in laboratory strains [21]. Our results further support and explain how the presence of external layers influence mass transfer in cyanobacteria. On the one hand, cyanobacteria cells are able to actively take up nutrients from the boundary layer adjacent to the cell, and the presence of a gelatinous coat offers an effective way to increase the prospects of encountering nutrient molecules in water [9]. On the other hand, gelatinous layers can simultaneously maintain a unique microenvironment, which allows cells to rapidly take up nutrients across the cell wall. This process creates an immediate environment, in which the nutrient concentration is more dilute than in the medium, further contributing to a beneficial inward diffusion gradient of nutrients from the medium to the gelatinous layers [13]. However, it is also apparent that the investment of external gelatinous layers is taxon-specific, and the thickness and texture of the layers are different, and the environmental response is varied [14]. Thus, it is possible that provision of a mucilaginous coat of slime offers a different balance in the intracellular proportions of carbon, nitrogen, and phosphorus, when compared to the gelatinous coat sheath. These proportions further conspicuously influence carbon or nutrient cycling in water environments [51].
With global climate changes, cyanobacterial blooms are predicted to expand. External layers of cyanobacteria can also function as a self-defense mechanism, to protect cells from climate changes. As proposed by Reynolds [13], external layer production originated as a mechanism for regulating the accumulation of photosynthate in cells, which is not released in solution. Margalef [52] observed that the sheath coating the cell can minimize unnecessary metabolic activity by slowing down diffusion. In this sense, it is possible to infer that the creation of a gelatinous layer around the cells might facilitate the regulation of mass uptake and loss, mediate interplay between photosynthesis and photosynthesis-coupled respiration, and stabilize cell activity during periods of temperature fluctuation. Overall, the characterization and function of external layers are essential to gathering information about the persistent dominance of cyanobacteria in freshwaters. Clearly, further work will be required to understand such adaptive benefits and mechanisms.

Conclusions
In conclusion, the external layer extraction methods employed in the present study can provide stripped strains, and had no statistical influence on their photosynthetic activity. Through NMT analysis, we show that external-layer-retaining strains have higher net O 2 effluxes than stripped strains, while the role of slime in NH 4 + absorption is limited compared with that of sheath. Our instantaneous temperature shock experiments suggested that external-layer-retaining strains have temperature fluctuation tolerance over short time scales. We also deduced that the external layers, especially the dense sheath, are essential for this adaptation, and may have a buffer function and ameliorating impact on the efficiency of photosynthesis and photosynthesis-coupled respiration. These findings provide key insights into the dominance of cyanobacteria during climate changes.