Effect of volcanic eruption on nutrients, light, and phytoplankton in oligotrophic lakes

Limnol. Oceanogr., 58(4), 2013, 1165–1175E 2013, by the Association for the Sciences of Limnology and Oceanography, Inc.
Effect of volcanic eruption on nutrients, light, and phytoplankton in oligotrophic lakes Beatriz E. Modenutti,1 Esteban G. Balseiro,1,* James J. Elser,2 Marcela Bastidas Navarro,1Florencia Cuassolo,1 Cecilia Laspoumaderes,1 Maria S. Souza,1 and Vero´nica Dı´az Villanueva 1 1 Laboratorio de Limnologı´a, Instituto de Investigaciones en Biodiversidad y Medioambiente (INIBIOMA), Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas–Universidad Nacional del Comahue, Bariloche, Argentina 2 School of Life Sciences, Arizona State University, Tempe, Arizona Volcanic eruptions that shape the earth's surface can have major effect on ecosystems and, as natural experiments, can yield insights into ecological dynamics. On 04 June 2011, a mega-eruption in the Puyehuevolcanic complex (Chile) discharged massive amounts of ash and pumice. Using long-term data from five NorthAndean Patagonian lakes (Espejo, Correntoso, Nahuel Huapi, Gutie´rrez, and Mascardi) that received differinglevels of ash, we show that, in Lakes Espejo, Correntoso, and Nahuel Huapi, these inputs resulted in 1.5- to 8-foldincreases in total suspended solids, light extinction, phosphorus concentrations, and phytoplankton biomassrelative to pre-eruption conditions. Although ashes affected light scattering, the ultraviolet : photosyntheticallyactive radiation ratio remained , 0.30–0.35 in all the lakes and no changes were seen in dissolved organic carbonin the affected lakes post-eruption. Thus, no differential specific absorption of the different light wavelengthsoccurred due to ash input. The results of multiple regression analysis identified light extinction coefficient of PAR(KPAR) as the primary variable that was associated with variation in phytoplankton biomass (chlorophyll).
Furthermore, incubation experiments demonstrated significant effects of photoinhibition on phytoplanktongrowth in these lakes at ambient pre-eruption light intensities. Thus, we infer that increased phytoplanktonbiomass following the eruption likely reflects nutrient (phosphorus) loading and attenuation of excessive lightintensities.
Volcanic eruptions have shaped much of Earth's surface light intensities in the upper levels of the water column are over geological time, but they also, in the shorter term, known to reduce phytoplankton growth because of affect ecosystems at local, regional, and even global scales photoinhibition (Alderkamp et al. 2010; Gerla et al.
due to ejection and emission of gases, ashes, pumice, and 2011), consistent with the possibility that, in highly lava. Thus, eruptions present unique opportunities for transparent systems at least, increased light attenuation scientific discovery though such studies are often hindered by abiotic particles such as ash may positively affect by a lack of pre-eruption and post-eruption data that allow phytoplankton growth by reducing photoinhibition.
comprehensive assessment of their effects and the mecha- Despite widespread recognition that pelagic ecosystem nisms of those outcomes (Lindenmayer et al. 2010; Larson function reflects the joint effects of dynamic light and 2011). Past studies of the effect of eruptions on aquatic nutrient supplies modulated by water column physical ecosystems have emphasized fertilization by ash-borne structure and internal food web interactions (Sterner et al.
elements such as phosphorus and iron (Hamme et al.
1997; Falkowski and Raven 2007), the effect of light has 2010; Lin et al. 2011). Studies in marine environments have been largely neglected in studies regarding volcanic shown that, after volcanic eruptions, the concentrations of eruption. The 04 June 2011 explosion of Puyehue-Cordo´n chlorophyll, as a proxy of phytoplankton biomass, increase Caulle (40u359S, 72u079W) in southern Chile (Fig. 1A) (Hamme et al. 2010; Lin et al. 2011). Paleolimnological provided a unique chance for assessing such dimensions, as evidence from a lake in Iceland also shows that, after a the event deposited massive amounts of ash into a set of volcanic eruption that deposited considerable amounts of nearby lakes in Argentine Patagonia that has been tephra, there was an increase in chlorophyll-derived extensively studied for 17 yr (Morris et al. 1995; Callieri pigments in sediments, indicating an increase in phyto- et al. 2007; Corno et al. 2009). These temperate Andean plankton biomass following volcanic ash deposition lakes (located in North-Patagonia around 41uS) are (Einarsson et al. 1993). However, increased concentrations characterized by high transparency and high ultraviolet of suspended particles in the water column, such as radiation (UVR) penetration (Morris et al. 1995), where volcano-derived ashes, also increase light scattering and planktonic organisms living in surface waters are chroni- so decrease light penetration (Kirk 1994). In many cally exposed to high light intensity and irradiation at situations such shading would be expected to negatively damaging wavelengths (Modenutti et al. 2004, 2005). Such affect phytoplankton growth by reducing photosynthesis to an eruption presents not only an opportunity to evaluate levels where it does not exceed respiration (Huisman 1999; how volcanic eruptions affect lakes but also serves as a Huisman et al. 2002). However, in extremely transparent ‘‘natural experiment'' to test the roles of nutrient and light oligotrophic and ultraoligotrophic aquatic systems, high in the ecological functioning of large pelagic ecosystemsthat cannot otherwise be experimentally manipulated. To * Corresponding author: [email protected] assess these effects, we documented optical, chemical, and

Modenutti et al.
(A) Satellite view of the eruption in the Puyehue-Cordo´n Caulle volcanic complex at 18:45 h Universal Time Coordinated (UTC) on 04 June 2011 (Source: http://earthobservatory.
nasa.gov). (B) The five main study lakes in relation to the eruption site; the silhouette indicatesthe cloud seen in (A). 1 5 Espejo; 2 5 Correntoso; 3 5 Nahuel Huapi (NW sampling siteindicated); 4 5 Gutie´rrez; 5 5 Mascardi.
biological properties in five lakes receiving different levels water column using a PUV500B submersible radiometer of ash input and compared these to values from pre- (Biospherical InstrumentsTM). The profiler was lowered eruption data. To test for the mechanisms involved, we also at 0.2 m s21 and, at this descent rate, the temperature performed incubation experiments for two lakes, one resolution was . 0.1uC. The light profile included strongly affected by the eruption (Lake Espejo) and one ultraviolet (UV) bands (305, 320, 340, and 380 nm) as well relatively unaffected (Lake Mascardi), assessing the relative as photosynthetically active radiation (PAR; 400–700 nm).
effects of overall light intensity, light quality (e.g., UVR), Water samples were taken using a closing sampler at 0, 10, and nutrients on phytoplankton growth rate.
20, 30, and 45 m depth in the water column. Water sampleswere carried to the laboratory in thermally isolated containers within 3 h after sampling, and processedimmediately after arrival to the laboratory. A volume of Lake sampling—The field study comparing pre- and 200 mL was immediately fixed with acid Lugol's solution post-eruption data was carried out in five lakes: Espejo, for phytoplankton enumeration. A volume of 50 mL of Correntoso, Nahuel Huapi, Gutie´rrez, and Mascardi lake water was collected for enumeration of autotrophic (Fig. 1B; Table 1), which are part of the glacial lakes picoplankton in sterile tubes and fixed with 0.2 mm filtered district of the North-Patagonian Andes. Lakes Espejo and formaldehyde buffered with 0.1 mol L21 sodium cacodylate Correntoso are closer to the eruption and were subjected to (final concentration 2% vol : vol), stored in darkness at 4uC, considerable ashfall during the initial explosion and via and processed within 2 weeks (Callieri and Stockner 2002).
subsequent runoff (Fig. 1). Lake Nahuel Huapi is a large, For chlorophyll determinations, a volume of 200 mL from morphologically complex lake whose northwestern arm each sampling depth was filtered onto Whatman GF/F (hereafter: NW Nahuel Huapi) received direct surface filters; in a more restricted set of samples, we also filtered inputs of ash as well as inputs from inflow streams (Fig. 1).
onto 0.2 mm pore-size polycarbonate filters (Nuclepore).
Lakes Gutie´rrez and Mascardi received little ash and allow All filters were frozen until extraction.
us to assess variability in the absence of eruption (Fig. 1).
Throughout the post-eruption study period, lakes were Laboratory determinations—The readily available P sampled at a standard deep-water station at approximately content of ashes freshly collected at Bariloche during the 2–3 week intervals during the austral growing season initial ash fall was measured by suspending 1 g of ashes in (October 2011–March 2012). Vertical temperature and light 50 mL of MilliQTM water for 1 h and then evaluating profile measurements were taken in the upper 50 m of the soluble reactive phosphorus by the molybdate reaction Location and morphometric characteristics of the studied North-Patagonian Andean lakes. Abbreviations: masl, meters above sea level; Zmax, maximum depth of the lake.
Effect of volcanic eruption in lakes (APHA 2005). For each sampling depth in routine with epilimnetic lake water that was diluted 80% with sampling, measurements of total phosphorus (TP; unfil- filtered (GF/F glass-fiber filter) lake water to minimize tered lake water) and total dissolved phosphorus (TDP; grazing and allow for growth in the unenriched treatment.
Whatman GF/F–filtered lake water) concentrations were During the experiment, temperature was maintained at , performed using the molybdate reaction after persulfate 18 6 1uC (similar to surface lake temperatures) and digestion of the sample. Dissolved organic carbon (DOC) incident solar irradiance was monitored with a photo- was determined on filtered lake water (pre-combusted GF/ radiometer. Experiments were run for 72 h. Chl a was F Whatman filters) using a ShimadzuTM total organic measured at the start and end of the experiment.
carbon analyzer.
The Lake Mascardi experiment was completed from 25 Chlorophyll a (Chl a) was extracted in hot ethanol to 28 January 2012, whereas the experiment for Lake following Nusch (1980) and determined by fluorometric Espejo took place from 02 to 05 February 2012. During this analysis (Turner DesignsTM 10AU fluorometer) with acid period the skies were generally cloudless; maximum daily correction. Total suspended solids (TSS) were estimated by incident irradiance for PAR was 2900 mmol photons filtering 1 liter of lake water onto pre-weighed GF/F filters m22 s21, whereas the daily irradiance dose was that were then dried at 60uC for 48 h and reweighed. The 92 mol m22 for the 14 h day length. Maximum incident two filter types (GF/F and 0.2 mm Nucleopore) gave very intensities for UVR at 320 and 340 nm were 44 and similar values in Chl a (no significant differences). Finally, 74 mW cm22 nm21, respectively, while their respective daily filters for seston C analysis were prepared by filtering lake doses were 11 kJ m22 and 20 kJ m22.
water onto pre-combusted Whatman GF/F glass-fiberfilters; these were dried and held in a desiccator until later Calculations—For comparing pre- and post-eruption analysis on a Thermo Finnigan EA1112 elemental analyzer.
data for TP and Chl a, the multiple depths in the water Autotrophic picoplankton, mainly picocyanobacteria, column were averaged. Note that because our sampling were counted on black polycarbonate filters (Poretics, depths are approximately evenly spaced across the upper 0.2 mm pore size) by autofluorescence of phycoerythrin water layers and encompassed the deep chlorophyll using an Olympus BH 50 epifluorescence microscope fitted maximum, average chlorophyll concentration (per unit with blue excitation cube (U-MWB) and green excitation volume) is essentially equivalent to integrated chlorophyll cube (U-MWG) light filters. Cells were counted using an concentration from 0–45 m (per unit area). Long-term data image analysis system (Image ProPlus; Media Cybernetics).
for all parameters were then analyzed by calculating an Phytoplankton were enumerated to genus and/or species overall, pre-eruption, average for each parameter for each level using 50 mL settling chambers with an inverted 4 week interval during the October–February period. All microscope. The four dominant phytoplankton types individual data points were then normalized to that average enumerated and analyzed reflect the species typically found (by dividing) for each 4 week data bin; thus, a value in an in Andean lakes over the past four decades (Thomasson interval that does not differ from the overall pre-eruption 1963; Callieri et al. 2007). These are Chrysochromulina dynamics would have a value of 1. We then calculated parva (Haptophyceae), Rhodomonas lacustris (Cryptophy- averages and 95% confidence intervals for these normal- ceae), Aulacoseira granulata and Cyclotella stelligera ized, pre-eruption data for each data bin. To plot the data (Bacillariophyceae), and Gymnodimium paradoxum and on a common graph for more than one lake (e.g., for the Gymnodimium varians (Dinophyceae). The presence of two unaffected lakes, Mascardi and Gutie´rrez), we plotted Tabellaria flocculosa in the post-eruption samples was the maximum confidence limits for each 4 week data confirmed by microscopic analysis (Koppen 1975).
period. We normalized each post-eruption data point to itspre-eruption mean for the appropriate temporal bin and Incubation experiments—To test the dependence of plotted those data along with the normalized pre-eruption phytoplankton growth rate on nutrients and light in both confidence intervals; statistically significant deviations Lake Mascardi (unaffected by ash inputs) and Lake Espejo from historical dynamics were assessed by determining if (affected by ash), we ran two outdoor incubation experi- post-eruption values were outside of these 95% confidence ments. A full 2 (unenriched, enriched) 3 2 (full solar bands. The dependence of Chl a on light and TP radiation [FSR]; PAR) 3 5 (100%, 75%, 33%, 11%, and concentration was assessed with forward and backward 3% of solar radiation) factorial design (five replicates per treatment) was carried out in 500 mL polypropylene bags In addition, we analyzed light transparency and phyto- arranged in a 1 m3 water bath. Nutrients were enriched plankton composition of the pre- and post-eruption data using the freshwater culture medium COMBO (Kilham et set for the lake thermal stratified period (December– al. 1998) to increase P by 15 mg L21; note that COMBO February). Light attenuation coefficients (K) were estimat- contains N (as NO3; N : P 5 44 by mass) as well as a mix of ed from the vertical light profiles as the slope of loge- trace metals. Light was manipulated in two different ways, transformed irradiance data with depth. To evaluate quality and quantity. The quality, or PAR treatment, was possible UV wavelength–specific changes relative to PAR produced using cut-off filters to remove wavelengths absorption due to suspended ashes, we also analyzed the shorter than 400 nm (CourtgardTM, CPFilms; Doyle et al.
ratio of the 1% depth of 320 nm UV relative to the 1% 2005). Light quantity was manipulated by reducing depth of PAR following Rose et al. (2009).
irradiance with increasing layers of shade screens to achieve For calculation of the mean light intensity in the mixed the five levels of incident irradiance. Each bag was filled layer (Im; Sterner et al. 1997) for all sampling dates, we used Modenutti et al.
2011 of , 14 mg L21 in Lakes Espejo, Correntoso, andNW Nahuel Huapi, relative to values typically closer to ,0.5 mg L21 (maximum: , 1 mg L21) in these lakes beforeeruption and in Lakes Gutie´rrez and Mascardi both beforeand after the eruption. During summer 2011–2012, TSSlevels in the affected lakes (Espejo, Correntoso, and NahuelHuapi) were 2–8 times higher than typical pre-eruptionlevels (Fig. 2). Ash inputs had major effects on variousother limnological variables in the affected lakes (Fig. 3).
Prior to eruption, the five lakes were very transparent, withlight extinction coefficients (KPAR) generally 0.10–0.15 m21(Table 2); however, post-eruption data showed that KPARincreased 1.5- to 2.5-fold after the eruption in LakesEspejo, Correntoso, and Nahuel Huapi, while remaining atthe same historical levels in Mascardi and Gutie´rrez(Fig. 3A).
In two of the affected lakes (Lakes Espejo and Post-eruption dynamics of TSS in the affected lakes, Correntoso), TP concentrations increased up to . 3-fold relative to typical pre-eruption values that never exceeded 1 mg L21 post-eruption (Table 2; Fig. 3B; increasing from values of for a 15-yr record (horizontal line).
, 2.0–3.4 mg P L21 to , 4.2–8.4 mg P L21), a result thatlikely reflects direct contributions of suspended ash, as light extinction coefficients (K chemical analysis of fresh ash collected during the 04 June PAR) obtained as described above along with estimates of the depth of the mixing layer 2011 event indicated an available P content of 0.009% by mass. There was no such increase in TP concentrations in m) from the vertical temperature profile obtained at the Lake Nahuel Huapi after the eruption (Table 2; Fig. 3B; m was determined as the depth above the temperature discontinuity (temperature difference . 1uC from 3.9–4.5 mg P L21 to 5.6 mg P L21). The unaffected m21) identified by direct inspection of the continuous Lakes Gutie´rrez and Mascardi retained typical TP concen- vertical temperature profiles obtained with the PUV trations of 1.6–4.3 mg P L21 (Table 2; Fig. 3B). TDP radiometer. Since the lakes were generally sampled on showed a similar pattern to that of TP, including an calm days in this windy region, small thermal gradients absence of nutriclines in the vertical profiles. In two of the (, 1uC) due to temporary diurnal microstratification were affected lakes (Espejo and Correntoso), TDP increased ignored in estimating Z from 1.1–2.3 mg P L21 before the eruption to , 4.0–4.5 mg m on a given day; note that most of our sampling was relatively early in the morning so that P L21 after the eruption, whereas TDP in Lake Nahuel such temporary stratification events were uncommon in the Huapi did not change appreciably (2.0–2.8 mg P L21 before to 3.5 mg P L21 after). The unaffected Lakes Gutie´rrez and Mascardi always had very low values of TDP (1.0–2.4 mg m was then estimated as in previous work (Sterner et al.
Phytoplankton biomass as indicated by Chl a increased strongly after the eruption in Lakes Espejo, Correntoso, and NW Nahuel Huapi, with Chl a as much as four times higher (Fig. 3C; Table 2), increasing from concentrations Note that calculation of Im yields a value for the average typically , 0.6 mg L21 for Correntoso and Espejo and light experienced across the entire mixed layer as a fraction , 0.9 mg L21 for NW Nahuel Huapi. However, Chl a in the of surface irradiance; this Im varies from 0 to 1.
unaffected lakes (Gutie´rrez, Mascardi) did not deviate The incubation experiment data were analyzed as growth from pre-eruption dynamics, maintaining concentrations rates based on Chl a data. Growth rate (GR) for each , 0.85 mg L21 (Fig. 3C; Table 2).
replicate for the 3-d incubation was calculated as: To assess the relative contributions of light attenuation and nutrient loading to observed increases in lake phytoplankton biomass, we performed stepwise multiple regression. Both forward and backward algorithms relating where Chl ai and Chl af are the initial and final Chl a to KPAR and TP for all data from the five lakes concentrations of Chl a. GR data were then analyzed with indicated that TP was eliminated and only KPAR remained a three-way factorial analysis of variance (ANOVA).
in the final model (r2 5 0.41) with a highly significant (p , Normality and homoscedasticity were confirmed prior to 0.001) and positive association with Chl a. The same the ANOVA; the data did not require transformation.
overall result was obtained when the analysis was confinedto only the affected lakes following the eruption, during which time KPAR would be dominated by light extinctioncontributed by ash particles. In addition, a similar result In situ dynamics—Immediately after the eruption, was obtained if TDP rather than TP was used in the concentrations of TSS increased, reaching levels in July multiple regression analysis; only KPAR was retained.
Effect of volcanic eruption in lakes Post-eruption changes in (A) optical (light extinction coefficient [KPAR]), (B) chemical (total phosphorus concentration [TP]), and (C) biological (phytoplankton biomass asindexed by chlorophyll a concentration [Chl a]) properties in the lakes, normalized to pre-eruption values from , 17 yr of monitoring (data in Table 2). Thus, a value of 1 indicates nochange relative to the corresponding pre-eruption interval. Dotted lines indicate 95% confidencelimits for each month's pre-eruption data.
More detailed analysis of vertical profiles shows that increased only , 1.5-fold during the early part of the light penetration changed up to 2.7-fold in the affected growing season in the post-eruption period in this lake lakes, particularly in Lakes Espejo and Correntoso (Fig. 4).
(Fig. 3A). In addition, we note that Lake Nahuel Huapi As an average for the summer stratification period, in Lake has a considerably deeper thermocline compared with the Espejo the depth of 1% light penetration (Z1%) for PAR other two affected lakes; this causes a different physical decreased from , 40 m to 20 m (Fig. 4A,B) and in Lake structure in which the euphotic zone is coincident with or Correntoso from 42 m to 25 m (Fig. 4C,D). Because near the upper limit of the mixolimnion (Fig. 4E,F).
thermal structure of the lakes did not change after the The ratio between UV wavelength–specific changes eruption, these increases in light attenuation also caused relative to PAR showed that the ash inputs affected all decreases in the mean irradiance of the mixing layer (Im), wavelengths equally and thus no shift was found before and from . 60% of surface irradiance in Espejo and after the eruption, with the UV : PAR ratio remaining , Correntoso before the eruption to , 30% afterwards.
0.30–0.35 in all the lakes. This result likely reflects the lack Notably, average Im in Lake Nahuel Huapi over the of change in DOC in the affected lakes post-eruption; pre- summer thermal stratification period did not decrease to eruption DOC concentration did not exceed 0.6 mg L21 the same degree as in the other two lakes because mean Kd (Morris et al. 1995) and these values were maintained after Pre-eruption mean values of chlorophyll (Chl a; mg L21), total dissolved phosphorus (TP; mg L21), and light extinction coefficient (KPAR; m21) for each month for each lake, used for standardization in Fig. 3. Data are from 1994–2010.
Modenutti et al.
Box plot of the chlorophyll a : sestonic C (Chl a : C) ratio (mg : mg) for the study lakes during 2011–2012 samplingperiod. Box limits indicate 25th and 75th percentiles; horizontalline in the box represents the median.
average length 5 3.78 6 0.5 mm) with increased abundanc-es of both towards 45 m depth (Fig. 6A,B; see pre-eruptiongraphs). However, examination of post-eruption samplesindicated that, whereas picocyanobacteria remained thedominant component of the phytoplankton (Fig. 6B; seepost-eruption graphs), there was a noticeable change in thestructure of the phytoplankton community (Fig. 6A; seepost-eruption graphs). In Lakes Correntoso and Espejo weobserved an increase in the abundance of Cryptophyceae,in particular of Rhodomonas lacustris (cell average length 58.7 6 0.7 mm). In Lake Nahuel Huapi we noted a decreaseof C. parva with a concomitant increase in the abundanceof R. lacustris and diatoms (Tabellaria flocculosa andAulacoseira granulata; Fig. 6A; see post-eruption graphs).
Depth profiles for mean percent irradiance for 305, After the volcanic event we observed that the increase in 320, 340, and 380 nm UV, and PAR and temperature for (A,C,E) phytoplankton biomass occurred both in the deep chloro- pre-eruption and (B,D,F) post-eruption periods. Pre-eruption phyll maxima (DCM) and in the mixing layer, and that the data are average of 2000–2008 summer (Dec–Feb). Post-eruption mean depth of the maximum cell abundance of both data are averages of December 2011–February 2012 sampling phytoplankton components moved upwards to around 20 m dates. References: A and B: Lake Espejo, C and D: Lake depth, especially in Espejo and Correntoso (Fig. 6).
Correntoso, E and F: Lake Nahuel Huapi.
Incubation experiments—Our incubation experiments the eruption (Espejo 5 0.52 6 0.15 mg L21, Correntoso 5 showed that, while nutrient fertilization modestly increased 0.51 6 0.14 mg L21, and Nahuel Huapi 5 0.51 6 phytoplankton growth, consistent with a role for ash-borne 0.11 mg L21).
nutrients in stimulating post-eruption chlorophyll concen- The Chl a : C ratio (mg : mg) averaged 10–12 in the trations in the affected lakes, there was also a large negative affected lakes as well as in the unaffected Lakes Gutie´rrez effect of overall light intensity on phytoplankton GR as and Mascardi (Fig. 5). Furthermore, we found no signif- well as an effect of UV removal (Fig. 7). GR declined at icant differences in Chl a : C between lakes (ANOVA, F4,35 relatively low levels, when FSR intensity exceeded 10% of 5 1.43, p 5 0.24), although Lake Nahuel Huapi had more incident irradiance (or even 3% for the unenriched variable Chl a : C ratios (Fig. 5).
treatment in Lake Mascardi). Similar, though slightly more Previous quantitative data on summer phytoplankton modest, negative effects of light intensity were obtained taxa in Lakes Espejo, Correntoso, and Nahuel Huapi when UVR was removed. Notably, the difference between indicated dominance by picocyanobacteria and the nano- FSR and PAR treatments was only present in medium to flagellate Chrysochromulina parva (Haptophyceae; cell high light intensities, while at low light intensities there was Effect of volcanic eruption in lakes Vertical profiles of phytoplankton in Lakes Espejo, Correntoso, and NW Nahuel Huapi pre-eruption and post-eruption. (A) Phytoplankton composition and cell abundance, (B)picocyanobacteria cell abundance. Pre-eruption data are average of 2000–2008 summer (Dec–Feb). Post-eruption data are averages of December 2011–February 2012 sampling dates.
little effect of UVR removal, especially in the Espejo eruption, from , 62% (brighter than Lake Mascardi) to , experiment. For both lakes, ANOVA indicated highly 30%. This change was nearly entirely due to the increase in significant main effects of light quality (p , 0.0001), light KPAR post-eruption, as average mixing depth did not intensity (p , 0.0001), and nutrients (p , 0.0001) on GR.
change appreciably following the eruption (Fig. 4A,B). In We found statistically significant light intensity 3 nutrient the Lake Espejo experiment (Fig. 7B), a decline in Im from interactions in both experiments (p 5 0.006 in Espejo and 62% to 30% Im corresponds to a shift from light conditions p 5 0.024 in Mascardi) but only the Espejo experiment had that strongly inhibited phytoplankton growth (regardless of significant two-way interactions involving light quality, nutrients or UV) in the experiment to those at which GR with intensity (p , 0.0001) and with nutrients (p 5 0.026)).
was near zero or positive.
No significant (p . 0.05) three-way interactions wereobserved.
To interpret these results in light of the Puyehue eruption, we assessed the shading stimulation of phyto- Overall, lake dynamics after eruption indicate that plankton GR in light of the observed values of average increased phytoplankton biomass was likely due to relative mixed-layer irradiance (Im expressed as a percent- combination of an increase in P supplies together with a age of surface irradiance) before and after the eruption. Im lowering of light intensities causes by suspended ashes. An was , 42% in Lake Mascardi and did not change after the indication that mechanisms other than nutrient loading eruption. In the Lake Mascardi experiment (Fig. 7A), alone appear to be involved was provided by the phytoplankton had near-zero or slightly negative growth observation that Chl a increased considerably in October rates at this Im, even with nutrient fertilization and UVR and November in NW Nahuel Huapi after the eruption, removal. In Lake Espejo, Im declined markedly after the despite no apparent change in P (Fig. 3; see October and Modenutti et al.
and because of its location and the predominance ofwestern winds, has a considerable fetch (Fig. 1) that resultsin a very deep thermocline (Fig. 4E,F). Nevertheless,almost the whole epilimnion is exposed to UV-A whilethe upper 40% of the epilimnion is also exposed to UV-B(305 nm; Fig. 4). Indeed, in these North-PatagonianAndean lakes DOC concentrations are very low (Morriset al. 1995) and the lakes have elevated UV : PAR ratios (,0.35), indicating that light attenuation is not due to DOCbut instead is dominated by attenuation by suspendedparticles (Rose et al. 2009). The fact that suspended ashesincreased KPAR and light scattering at all wavelengths (nochanges in UV : PAR ratio were observed) implies anoverall reduction in total solar radiation received as well asamelioration of UVR exposure (Fig. 4).
The inference that increased post-eruption chlorophyll at least partially reflects lower light intensities is supported bythe results of multiple regression analysis, which identifiedKPAR as the variable that is primarily associated withvariation in Chl a, both across all lakes throughout thestudy period as well as just in the affected lakes during thepost-eruption interval. Importantly, this correlation anal-ysis is bolstered by the experimental results (Fig. 7):nutrient enrichment had only modest effect on phytoplank-ton growth compared with the large positive effect oflowering solar radiation. We note that the modestresponses to nutrient enrichment we observed in theseexperiments are unlikely to reflect possible nutrient releasefrom cells damaged during preparation of filtered lakewater for the 80% experimental dilution because measure-ments of soluble (i.e., filtered) reactive phosphorus samplestypically are below our limit of detection of 1 mg L21,considerably lower than the 15 mg L21 experimental Penrichment. Removal of UVR also had a positive effect on Results of incubation experiments testing the effects GR, an understandable outcome given that ultraviolet of overall light intensity (UVR+), removal of ultraviolet radiation radiation is known to damage phytoplankton and reduce (UVR2), and nutrients on growth rate of phytoplankton from primary production (Holm-Hansen et al. 1993; Neale et al.
(A) Lake Mascardi (relatively unaffected by the eruption) and (B) 1998a). When we removed this damaging wavelength in our Lake Espejo (strongly affected by the eruption). Error bars on experiments, the overall photoinhibition effect was reduced each symbol indicate 6 1 standard error. On each figure, vertical but not eliminated. This result suggests that PAR itself is lines indicate the historical 17 yr pre-eruption average mixed-layer too high and sufficient on its own to induce photoinhibition light intensity (Im, as a percentage of incident) as well as theaverage post-eruption value for 2011. For Lake Mascardi, the pre- and that this effect is not counteracted by nutrient and post-eruption values of I m were essentially identical, but the horizontal arrow in panel B indicates the post-eruption shift seen Our inference that excess irradiance is an important in Lake Espejo due to shading by suspended ash.
ecological factor in these lakes clarifies the observeddynamics in Lake Nahuel Huapi in spring and Lakes November samples). Furthermore, Chl a decreased in early Espejo and Correntoso in early summer, where phyto- summer (December to January) in Lakes Espejo and plankton biomass changed significantly despite no appar- Correntoso with no decrease in P (Fig. 3A) but in concert ent change in nutrient levels (Fig. 3) but in concert with with declining KPAR (increasing light). One possible shifts in light extinction. Consistent with this interpreta- mechanism for these changes in Chl a despite no change tion, previous studies in the same lakes have shown that in TP is that suspended ash ameliorated exposure of primary production at depths corresponding to 10% of phytoplankton to excessive solar radiation, which previous surface irradiance is 5–10 times higher than that in surface studies have shown to be damaging in these highly layers (Modenutti et al. 2004; Callieri et al. 2007), despite transparent lakes (Morris et al. 1995; Modenutti et al.
vertical uniformity in nutrient concentrations.
2004; Villafan˜e et al. 2004). Previous to the eruption event, Extrapolation of our experimental results to lake wind-sheltered lakes like Espejo and Correntoso had their conditions is complicated by the fact that we used a static whole epilimnia illuminated and exposed to high PAR as incubation but, under natural conditions, phytoplankton well as hazardous UVR (including UV-B and UV-A; cells in a lake's mixing layer are continuously brought in Fig. 4). Lake Nahuel Huapi is the largest lake in the area, and out of near-surface layers where solar radiation is high Effect of volcanic eruption in lakes and induces photoinhibition (Neale et al. 1998b; Modenutti volcanic activity. However, diatoms would likely not be et al. 2005). Relevant to such mechanisms, our experimen- supported in Espejo and Correntoso because their shal- tal results suggest a stronger negative effect of light on lower mixing layers have low turbulent diffusivity com- phytoplankton growth rate at irradiances exceeding Im pared to Nahuel Huapi (Huisman et al. 2004).
than below (Fig. 6). This apparent nonlinearity suggests Picocyanobacteria remained the dominant component in that, compared to cells held at a constant value of Im (as in the phytoplankton based on cell abundance both in pre- our incubation experiment), overall photoinhibitory effects and post-eruption periods (Fig. 6B). Picocyanobacteria would be larger for circulating cells experiencing light . Im have generally more relative abundance when nutrient for part of the day and light , Im for part of the day.
concentrations are low (Callieri et al. 2007). For all three Considering the differences in Im between Lakes Espejo and affected lakes, the abundance of picocyanobacteria de- Correntoso vs. Lake Nahuel Huapi, it seems that our static clined after the eruption. Further, and consistent with a experiments could be either under- or overestimating decrease in nutrient limitation, the abundance of larger photoinhibitory effects on phytoplankton experiencing in nanoplankton increased. Picocyanobacteria have been situ Im. Nevertheless, considering the in situ dynamics reported to perform well under low-intensity, green light following the volcanic natural experiment and our exper- because of the presence of phycoerythrin (Stomp et al.
imental results together, photoinhibition emerges as a 2007). This helps explain increased abundances of phyco- candidate factor contributing to dynamics of primary erythrin-rich cells previously observed in the DCM of deep production and phytoplankton biomass under normal ultraoligotrophic Patagonian lakes (Callieri et al. 2007), (pre-eruption) conditions in these lakes and suggests that where blue-green light prevails (Pe´rez et al. 2002). Since no excessive irradiance plays a similar role in other water post-eruption changes were observed in absorption of columns of high optical clarity.
different wavelengths (Fig. 4), dim, green light conditions This inference is strengthened by previous studies in after the eruption occurred at a shallower depth (see these lakes, which also provide insight into how phyto- increase around 20 m depth in Fig. 6B post-eruption plankton vertical distributions shifted in response to graphs). Thus, as the same light quality was achieved at the eruption. Phytoplankton in clear North Andean- shallower depth, picocyanobacteria moved upwards in the Patagonian lakes often develop DCM, likely due to strong water column. These shifts are consistent with previous effects of photoinhibition (Modenutti et al. 2004). There- documentation of the highly variable dynamics of pico- fore, the possibility of a refuge against hazardous cyanobacteria in forming DCM in the metalimnion or the wavelengths in deep layers is important in these extremely hypolimnion (Callieri et al. 2007).
clear lakes. Previous to the volcanic event, these lakes Consistent with an overall view that high light intensities regularly exhibited DCM, either in the hypolimnion are an important ecological factor in these lakes are the low (Espejo and Correntoso) or in the metalimnion (Nahuel Chl a : C ratios we observed, as low Chl a : C ratios are Huapi; Modenutti et al. in press). These DCM involve generally considered to be indicative of high irradiance mainly motile mixotrophic cells (Modenutti et al. 2004, in conditions (Geider et al. 1997). While differences in average press) that are able to move to deeper layers (Sommaruga Chl a : C ratios between lakes were not significant, Chl a : C and Psenner 1997), exploiting hypolimnetic levels of the ratios in Lake Nahuel Huapi were quite variable (Fig. 5).
euphotic zone. Noticeably, after the eruption event, This variability may be due to several factors. Since phytoplankton composition changed, with an increase in different taxonomic groups can differ considerably in the the flagellate R. lacustris (Cryptophyceae) in Lakes Espejo Chl a : C ratios (Chan 1980), one possible contributing and Correntoso and an upward shift in the depth of its factor is the greater variability in algal composition and cell maximum abundance. This latter change is consistent with size (including nanoflagellates and diatoms, from , 3 to a community actively maintaining its position at a desired 45 mm in cell length) in Nahuel Huapi than seen in the other light intensity. R. lacustris is a facultative mixotroph that is lakes. Another possible contributing factor for the extreme very common in less transparent lakes in the area (Balseiro variability in Chl a : C ratios in Nahuel Huapi is its deeper et al. 2004) and the flexibility of such mixotrophs has been thermocline, its complex lake morphometry, and its greater suggested as an adaptive advantage that allows them to wind exposure than in the other lakes. Together, these may dominate in plankton communities exposed to variation in result in the whole euphotic zone being included in the light (Laybourn-Parry et al. 1997). This change in flagellate mixing layer, leading to the possibility of a weak photo- species composition implies a change in cell size (domi- acclimation because turbulence may drag cells all along the nance from C. parva, cell average length 5 3.78 6 0.5 mm, light gradient (Geider et al. 1997) or sporadically move to R. lacustris, cell average length 5 8.7 6 0.7 mm), which them out of the DCM into the mixed layer.
may help in understanding the observed increase in Chl a Beyond documenting volcano effects on lake water concentrations despite no significant change in cell quality, our study suggests an unexpectedly contribution numbers or Chl a : C ratio. On the other hand, the increase of excess light in affecting phytoplankton growth and of the diatoms A. granulata and T. flocculosa in Lake production in these transparent Patagonian lakes. It has Nahuel Huapi and their shift towards deeper levels suggest recently been suggested that limiting light conditions are an that these nonmotile cells accumulated at the thermocline underappreciated factor in lake ecology (Karlsson et al.
because they cannot actively regulate their position in the 2009). However, the findings of Karlsson et al. are from water column (Cullen 1982). The increase in diatoms may small, highly colored lakes that, while numerous, do not be a result of more silica in the water column from the contain large volumes of surface freshwaters. In such Modenutti et al.
shallow lakes, colored dissolved organic matter greatly BEETON, A. M. 2002. Large freshwater lakes: Present state, trends, reduces overall light intensity and acts as the ‘‘ozone of the and future. Environ. Conserv. 29: 21 underwater world'' by selectively removing the shorter wavelengths of light (Williamson and Rose 2010). However, BOYCE, D. G., M. R. LEWIS, AND B. WORM. 2010. Global in these clear North Patagonia lakes, dissolved organic phytoplankton decline over the past century. Nature 466: matter concentration is extremely low (, 0.6 mg L21; Morris CALLIERI, C., B. MODENUTTI, C. QUEIMALIN˜OS, R. BERTONI, AND et al. 1995; Corno et al. 2009) and high light has been E. BALSEIRO. 2007. Production and biomass of picophyto- recognized as an important factor structuring lacustrine plankton and larger autotrophs in Andean ultraoligotrophic plankton (Modenutti et al. 2004, 2005). The high-transpar- lakes: Differences in light harvesting efficiency in deep layers.
ency, low-nutrient scenarios typical of our study lakes are Aquat. Ecol. 41: likely widespread in many large lakes worldwide; notably, ———, AND J. G. STOCKNER. 2002. Freshwater autotrophic large lakes contain the majority (68%) of global unfrozen picoplankton: A review. J. Limnol. 61: 1–14.
surface waters and provide a variety of key ecosystem CHAN, A. T. 1980. Comparative physiological study of marine services (Beeton 2002). Such high light scenarios may also be diatoms and dinoflagellates in relation to irradiance and cell common in many oceanic regions, leading us to suggest that size. II. Relationship between photosynthesis, growth, andcarbon/chlorophyll a ratio. J. Phycol. 16: increasing photoinhibition may be a factor contributing to recently reported declines in oceanic phytoplankton biomass CORNO, G., B. MODENUTTI, C. CALLIERI, E. BALSEIRO, R. BERTONI, linked to warming surface layers during the last 30 yr (Boyce AND E. CARAVATI. 2009. Bacterial diversity and morphology in et al. 2010; but see also McQuatters-Gollop et al. 2011). For deep ultraoligotrophic Andean lakes: The role of UVR on lakes, future light environments will depend strongly on vertical distribution. Limnol. Oceanogr. 54: 1098–111 possible changes in inputs of colored organic matter. While some early assessments suggested reduced inputs of colored CULLEN, J. J. 1982. The deep chlorophyll maximum: Comparing organic materials and thus brighter surface mixed layers vertical profiles of chlorophyll a. Can. J. Fish. Aquat. Sci 39: during coming decades (Magnuson et al. 1997), more recent studies highlight the possibility of increasing DOC inputs to DOYLE, S. A., J. E. SAROS, AND C. E. WILLIAMSON. 2005.
lakes due to temperature stimulation of DOC mobilization Interactive effects of temperature and nutrient limitation onthe response of alpine phytoplankton growth to ultraviolet from peatlands (Freeman et al. 2001; Hansson et al. 2012).
radiation. Limnol. Oceanogr. 50: According to the inferences we make here, such changes might result in increased phytoplankton production in EINARSSON, A., H. O ´ SKARSSON, AND H. HAFLIDASON. 1993.
highly transparent lakes by shielding phytoplankton from Stratigraphy of fossil pigments and Cladophora and its excessive irradiance. Regardless of the direction of such relationship with deposition of tephra in Lake My´vatn, DOC shifts, our work suggests that understanding the effects Iceland. J. Paleolimnol. 8: of shifts in DOC inputs to lakes will require assessment of FALKOWSKI, P. G., AND J. A. RAVEN. 2007. Aquatic photosynthe- how such changes might modulate photoinhibition of lake sis, 2nd ed. Blackwell Science.
ecosystem production and interactions with nutrient supply.
FREEMAN, C., C. D. EVANS, D. T. MONTEITH, B. REYNOLDS, AND Thus, our findings should motivate continued work on the N. FENNER. 2001. Export of organic carbon from peat soils.
Nature 412: role of optical conditions in affecting the functioning of the GEIDER, R. J., H. L. MACINTYRE, AND T. M. KANA. 1997.
pelagic ecosystems that are fundamental in global water Dynamic model of phytoplankton growth and acclimation: supplies and biogeochemical cycles.
Responses of the balanced growth rate and the chlorophylla : carbon ratio to light, nutrient-limitation and temperature.
Mar. Ecol. Prog. Ser. 148: J.J.E. was supported by the Fulbright Foundation and the GERLA, D. J., W. M. MOOIJ, AND J. HUISMAN. 2011. Photoinhibition National Science Foundation (DEB-0950179). We thank two and the assembly of light-limited phytoplankton communities.
anonymous reviewers for their constructive comments on this Oikos 120: 359–3 manuscript. Data were produced via grants to B.M. and E.B.
HAMME, R. C., AND OTHERS. 2010. Volcanic ash fuels anomalous from the Fondo para la Investigacio´n Cientı´fica y Tecnolo´gica plankton bloom in subarctic northeast Pacific. Geophys. Res.
(FONCyT, Argentina) and the National Geographic Society.
HANSSON, L.-A., AND OTHERS. 2012. Food-chain length alters community responses to global change in aquatic systems.
Nature Clim. Change 3: 228–233.
ALDERKAMP, A.-C., H. J. W. DE BAAR, R. J. W. VISSER, AND K. R.
HOLM-HANSEN, O., E. W. HELBLING, AND D. LUBIN. 1993.
ARRIGO. 2010. Can photoinhibition control phytoplankton Ultraviolet radiation in Antarctica: Inhibition of primary abundance in deeply mixed water columns of the Southern production. Photochem. Photobiol. 58: Ocean? Limnol. Oceanogr. 55: 1248–126 HUISMAN, J. 1999. Population dynamics of light-limited phyto- APHA. 2005. Standard methods for the examination of water and plankton: Microcosm experiments. Ecology 80: 202–210 wastewater. American Public Health Association, AWWA.
BALSEIRO, E. G., C. P. QUEIMALIN˜OS, AND B. E. MODENUTTI. 2004.
———, J. SHARPLES, J. M. STROOM, P. M. VISSER, W. E. A.
Grazing impact on autotrophic picoplankton in two south KARDINAAL, J. M. H. VERSPAGEN, AND B. SOMMEIJER. 2004.
andean lakes (Patagonia, Argentina) with different light : nutrient Changes in turbulent mixing shift competition for light ratios. Rev. Chil. Hist. Nat. 77: between phytoplankton species. Ecology 85: 2960–297 Effect of volcanic eruption in lakes ———, AND OTHERS. 2002. Principles of the light-limited MORRIS, D. P., AND OTHERS. 1995. The attenuation of solar UV chemostat: Theory and ecological applications. Antonie van radiation in lakes and the role of dissolved organic carbon.
Limnol. Oceanogr. 40: KARLSSON, J., P. BYSTROM, J. ASK, P. ASK, L. PERSSON, AND M.
JANSSON. 2009. Light limitation of nutrient-poor lake ecosys- NEALE, P. J., J. J. CULLEN, AND R. F. DAVIS. 1998a. Inhibition of tems. Nature 460: marine photosynthesis by ultraviolet radiation: Variable KILHAM, S. S., D. A. KREEGER, S. G. LYNN, C. E. GOULDEN, AND sensitivity of phytoplankton in the Weddell-Scotia Conflu- L. HERRERA. 1998. COMBO—A defined freshwater culture ence during the austral spring. Limnol. Oceanogr. 43: medium for algae and zooplankton. Hydrobiologia 377: ———, R. F. DAVIS, AND J. J. CULLEN. 1998b. Interactive effects KIRK, J. T. O. 1994. Light and photosynthesis in aquatic of ozone depletion and vertical mixing on photosynthesis of ecosystems. Cambridge Univ. Press.
Antarctic phytoplankton. Nature 392: KOPPEN, J. D. 1975. A morphological and taxonomic consider- ation of Tabellaria (Bacillariophyceae) from the northcentral NUSCH, E. A. 1980. Comparison of different methods for United States. J. Phycol. 11: 236–244.
chlorophyll and phaeopigment determination. Arch. Hydro- biol. Beih. Ergebn. Limnol. 14: 14–36.
ARSON, D. W. 2011. Science after the volcano blew: Research near Mount St. Helens proceeded despite bureaucratic PE´REZ, G. L., C. P. QUEIMALIN˜OS, AND B. E. MODENUTTI. 2002.
hurdles, limited funding and an extremely hazardous envi- Light climate and plankton in the deep chlorophyll maxima ronment. Am. Sci. 98: 324–333.
in North Patagonian Andean lakes. J. Plankton Res. 24: AYBOURN-PARRY, J., S. J. PERRISS, G. G. R. SEATON, AND J.
ROSE, K. C., C. E. WILLIAMSON, J. E. SAROS, R. SOMMARUGA, AND OHOZINSKI. 1997. A mixotrophic ciliate as a major contrib- utor to plankton photosynthesis in Australian lakes. Limnol.
J. M. FISCHER. 2009. Differences in UV transparency and thermal structure between alpine and subalpine lakes: Implications for organisms. Photochem. Photobiol. Sci. 8: IN, I. I., AND OTHERS. 2011. Fertilization potential of volcanic dust in the low-nutrient low-chlorophyll western North Pacific subtropical gyre: Satellite evidence and laboratory study.
SOMMARUGA, R., AND R. PSENNER. 1997. Ultraviolet radiation in a Global Biogeochem. Cycles high mountain lake of the Austrian Alps: Air and underwater measurements. Photochem. Photobiol.
LINDENMAYER, D. B., G. E. LIKENS, AND J. F. FRANKLIN. 2010.
Rapid responses to facilitate ecological discoveries from TERNER, R. W., J. J. ELSER, E. J. FEE, S. J. GUILDFORD, AND T. H.
major disturbances. Front. Ecol. Environ.
HRZANOWSKI. 1997. The light : nutrient ratio in lakes: The balance of energy and materials affects ecosystem structure and process. Am. Nat. 150: MAGNUSON, J. J., AND OTHERS. 1997. Potential effects of climate changes STOMP, M., J. HUISMAN, L. VOROS, F. R. PICK, M. LAAMANEN, on aquatic systems: Laurentian Great Lakes and Precambrian T. HAVERKAMP, AND L. J. STAL. 2007. Colourful coexistence Shield region. Hydrol. Processes 11: 825–87 of red and green picocyanobacteria in lakes and seas. Ecol.
MCQUATTERS-GOLLOP, A., AND OTHERS. 2011. Is there a decline in THOMASSON, K. 1963. Araucanian lakes. Acta Phytogeogr. Suec.
marine phytoplankton? Nature 472: E6–E7 47: 1–139.
VILLAFAN˜E, V. E., A. G. J. BUMA, P. BOELEN, AND E. W. HELBLING.
MODENUTTI, B., E. BALSEIRO, M. BASTIDAS NAVARRO, C. LASPOUMA- 2004. Solar UVR-induced DNA damage and inhibition of DERES, M. S. SOUZA, AND F. CUASSOLO. In press. Environmental photosynthesis in phytoplankton from Andean lakes of changes affecting light climate in oligotrophic mountain lakes: Argentina. Arch. Hydrobiol. 161: 245–26 The deep chlorophyll maxima as a sensitive variable. Aquat.
WILLIAMSON, C. E., AND K. C. ROSE. 2010. When UV meets fresh ———, ———, C. CALLIERI, C. QUEIMALIN˜OS, AND R. BERTONI.
water. Science 329: 2004. Increase in photosynthetic efficiency as a strategy ofplanktonic organisms exploiting deep lake layers. Freshw.
Biol. 49: Associate editor: John M. Melack ODENUTTI, B. E., E. G. BALSEIRO, C. CALLIERI, R. BERTONI, AND C. P. QUEIMALIN˜OS. 2005. Effect of UV-B and different PARintensities on the primary production of the mixotrophic Received: 22 December 2012 planktonic ciliate Stentor araucanus. Limnol. Oceanogr. 50: Accepted: 06 March 2013 Amended: 10 March 2013

Source: http://elserlab.asu.edu/pdf/Modenutti_2013.pdf