Layout 1 INTRODUCTION Many European lakes have been affected by a number of anthropogenic pressures, such as eutrophication, water level changes, toxic pollution and introduction of alien species. Mountain lakes, relatively far from heavy popu- lated areas, receive contaminants by long-range transport of air pollutants (Rogora et al., 2008; Poma et al., 2017). In the case of the Monticchio lakes, even if they are rela- tively far from human activities, and enclosed in Mt. Vul- ture caldera, strong human impact was due to the touristic development in the 1960s, when houses, restaurants and a cableway were built, at present in large part abandoned. In order to follow the temporal trend in the ecological quality of ecosystems, long-term monitoring using well defined protocols can be used, providing deep insight in the changes in ecosystem structure and function (Mora- bito et al., 2018). However, apart some rare exception (Minder 1938) systematic monitoring programmes are often started after ecosystem alteration was evident (Vol- lenweider et al., 1974; Keating and Dodd, 1975). Scattered morphological, physical and biological ob- servations are available for a number of lakes since the 1870s, concerning in particular the large lakes around the Alps (Marchetto, 1998), and the large volcanic lakes in Central Italy (Margaritora, 1992). Since the 1880s, a num- ber of studies also deals with the Monticchio lakes, and in this paper we evaluate the possibility to use them to infer the ecological history of these lakes during the peri- ods when they were affected by climate changes, by change of nutrient levels in water, mainly related to the tourist pressure. A list of the studies used for this historical reconstruc- tion is presented in the supplementary materials (Tab. S1). The older papers (Tata, 1778; Palmieri and Scacchi, 1852; De Giorgi, 1879; Cavanna, 1882; Marinelli, 1895; Vin- ciguerra, 1895; De Lorenzo, 1900; Casoria, 1901) only reported episodic information, frequently taken with rudi- mentary instruments. Later on, a number of studies are re- lated to one or few specific aspects of the lake water or biota, such as water chemistry (Cannicci, 1952; Squiccia- rini, 1974; Mongelli et al., 1975), phytoplankton (Can- nicci, 1952; Musacchio, 1981-1982), zooplankton (Ruffo and Stoch, 2005; Alfonso, 2008), or macrophytes (Venan- zoni et al., 2003; Azzella et al., 2010). However, four detailed limnological studies, covering most biological communities, were carried out. The first, on July 6th-8th, 1905, concerned lake morphometry, water chemistry, phytoplankton, zooplankton, macrophytes and phytobenthos and was reported by Forti and Trotter (1908) and Stegagno (1908). In 1991, a second study con- cerned lake morphometry, water chemistry, phytoplank- ton, zooplankton (Marano and D’Aprile, 1991). After these two short campaigns, a two year study was carried out in 2005-2007 by the Institute of Ecosystem Studies of the National Research Council (CNR-ISE) (Ceccanti et al., 2007), together with the Milano-Bicocca University, aiming to test the possibility to use the macrophytes to improve Lago Grande water quality and go into detail concerning water physical and chemical properties and ARTICLE The contrasting evolution of two volcanic lakes lying in the same caldera (Monticchio, Mt. Vulture, Italy) inferred from literature records Renato Spicciarelli,1* Aldo Marchetto2 1School of Agricultural, Forest, Food and Environmental Sciences (SAFE), University of Basilicata, Viale dell’Ateneo Lucano 10, 85100 Potenza; 2CNR Water Research Institute, Corso Tonolli 50, 28922 Verbania-Pallanza, Italy ABSTRACT Lago Piccolo and Lago Grande di Monticchio lie in the collapsed caldera of the volcanic structure of Mt. Vulture (Basilicata, Italy). In over two centuries, a number of studies on their water and on their submerged and riparian vegetation, were carried out, demonstrating an interesting biodiversity. The entire lake area, which is impacted by strong tourist pressure, is part of the “Monte Vulture” Special Area of Conservation (SAC IT9210210). The aim of this paper is to review the literature studies on these lakes, in order to identify the more suitable limnological parameters to infer the history of the trophic status of the two lakes. For this rea- son, we assess the current ecological status of the two lakes on the basis of physical, chemical and biological analyses deriving from two recent surveys carried out in 2005-2007 and in 2015, and compare these data with sparse, but relevant, historical records, in order to assess how human impacts affected both these lakes and to understand the differences in their present trophic status. Because of its peculiar water chemistry, Lago Piccolo is resulted in good and stable ecological conditions. On the contrary, water transparency of Lago Grande came out very low in summer, while total phosphorus and nitrogen concentration are proved high, leading to the persistence of critical environmental conditions in this lake, with high algal biomass and durable algal blooms in late summer, dominated by cyanobacteria. Finally, in absence of standard protocols and seasonal samplings, the macrophyte maximum growing depth should be considered the more reliable indicator of trophic status among those available for these specific lakes, being relatively independent from sampling methods and seasonal pattern. No n- co mm er cia l u se on ly Literature record of Monticchio lakes 45 phytoplankton and zooplankton communities. A last lim- nological campaign was carried out by the Basilicata Uni- versity (Spicciarelli and Mirauda, 2015), together with the Umbria and Marche Experimental Zooprophylactic Insti- tute and Umbria Regional Agency for Environmental Pro- tection. METHODS Study area Mount Vulture is a composite volcano located in the Basilicata region and formed by the superimposition of a number of volcanic edifices and affected by tectonic activ- ity (La Volpe et al., 1984). Its formation started 800-750 kyears ago (Principe, 2006). On the top of the mountain, at 650 m asl, a large caldera includes a tuff ring (Giannandrea et al., 2006) and two maars (Stoppa and Principe, 1997), hosting the two Monticchio lakes (Fig. 1), divided by a rock strip and connected by a narrow channel through which water flows from Lago Piccolo (LP) to Lago Grande (LG). LP is smaller (0.155 km2) and deeper (44 m) than LG (0.41 km2 and 40 m), but the shape of LP is closer to a cone, while LG presents a large shallow area, ca. 10 m deep (Fig. 2), so that the volume of LP (4.26 106 m3) is larger than the volume of LG (3.40 106 m3) (Spicciarelli and Mirauda, 2015). For LG, morphometric data obtained in 2015 compare well with those collected by Stegagno (1908). LG sedi- ment is mostly annually laminated and rich in tephras (Wulf et al., 2004, 2008, 2012; Schettler and Alberic, 2008), and represents one of the best sedimentary records for paleoclimatic studies in the Central Mediterranean (Watts, 1985; Allen et al., 1999; Brauer et al., 2007). The emerging watershed of the two Monticchio lakes comprises approximately the caldera, with a surface of ca. 4 km2, mainly forested. Beside precipitation in the catch- ment, Monticchio lakes, and in particular LP, are also fed by a heterogenous volcanic aquifer, with alternatively high and low permeability structures and several under- ground water divides (Celico and Summa, 2004). From the volcanic structure, carbon dioxide reaches the lakes, and the quantity of CO2 released compares with the two largest volcanoes producing CO2 (Gambardella, Fig. 1. Winter image of Mt. Vulture caldera with the two Monticchio lakes, during Lago Piccolo (on the right) partial overturn. No n- co mm er cia l u se on ly R. Spicciarelli and A. Marchetto46 2006), namely Mt. Etna (Sicily) and Popocatépetl (Mex- ico). For this reason, LP is considered at risk of limnic eruption (Chiodini et al., 1997, 2000; Cioni et al., 2006; Caracausi et al., 2009, 2013, 2015). In fact, a limnic erup- tion happened in 1810, and was described by Ferdinando Tortorella (1840) and quoted by another further visitor of Mt. Vulture (Gussone and Tenore, 1843; Malpica, 1847; Palmieri and Scacchi, 1852; Paci, 1853). Human activity in Mt. Vulture caldera dates back to the 4th-3rd century BCE, but a permanent settlement started in the 9th century, with the foundation of a Basilian monastery. Later a powerful Benedictine abbey took over. Anthropogenic pressure strongly increased in the 19th cen- tury, when forest surface was reduced because of tree cut- ting for using wood for railroad construction, and agriculture development, using lake water for irrigation. Human pressure increased again in the 1950s when tourism activities developed, with a strong impact on the Fig. 2. Bathymetric maps of Lago Grande (left) and Lago Piccolo (right): top, in 1905 (Stegagno, 1908); bottom, in 2015. Reproduced from Spicciarelli and Mirauda, UE- P.O. FEP 2007-2013 Regione Basilicata, 2015, with permission. No n- co mm er cia l u se on ly Literature record of Monticchio lakes 47 ecosystems. In 1963, an endemic moth (Brahmaea eu- ropaea Hartig) living in the Vulture caldera and in the close Grotticelle wood was described, mainly feeding on narrow-leaved ash (Fraxinus angustifolia subsp. oxycarpa (Willd.)) (Spicciarelli, 2004, 2015, 2018a, 2018b). It is considered a Miocene relict by some authors. For this rea- son, the Nature Reserve of “Grotticelle” was set in 1971, today included in the Special Area of Conservation (SAC) “Grotticelle di Monticchio” (Spicciarelli, 2013). At present, the intensity of the tourism activities is lower, and natural vegetation is recovering, but the filling of the LG outlet in 2011 caused the increase of lake level, damaging ash forest around the lake. RESULTS Temperature and transparency Both lakes show a clear stratification in summer, with the thermocline located between 5 and 10 m, while the temperature is homogenous in winter (Fig. 3). However, in LP there is an increase in temperature (up to 2.5°C) below 20 m, due to the presence of underwater springs. In the western part of LG, during summer and autumn the temperature between 4 and 7 m is 8-12°C lower than in the rest of the lake (Spicciarelli and Mirauda, 2015), prob- ably as an effect of underwater springs. Fig. 3. Profiles of water temperature in Lago Piccolo (left) and Lago Grande (right) in April, June, September, December 2005 September data for LP were not available. Redrawn from Ceccanti et al., CNR-ISE 2007, with permission. Fig. 4. Profiles of water pH in Lago Piccolo (left) and Lago Grande (right) in April, June, September, December 2005. Redrawn from Ceccanti et al., CNR-ISE 2007, with permission. No n- co mm er cia l u se on ly R. Spicciarelli and A. Marchetto48 During the recent studies, LG appeared warm monom- ictic, with full overturn in winter, as the volcanic lakes in central Italy (Margaritora, 1992). The same thermic regime was found by Mongelli (1964, 1975), Squicciarini (1974), and Nicolosi et al. (2010). However, LG may freeze in winter, behaving as a dimictic lake (Schettler and Albéric, 2008). On the contrary, in spite of the ho- moeothermic conditions reached in winter, LP is meromictic (see below). Several authors reported Secchi disk depth measured in the Monticchio lakes (Tab. 1). Comparing data from the end of the 19th century and the last decades, water transparency decreased in LG and increased in LP. Dissolved gases and water pH Underwater sources feed both Monticchio lakes in car- bon dioxide. In particular, in LP CO2 concentration as high as 23 to 41 mmol L–1 were measured below the 20 m depth. In LG, values of 17-21 mmol L–1 were measured below the 20 m depth, with a maximum of 31 mmol L–1 close to the lake bottom in September and October 2008 (Caracausi et al., 2013). Such high CO2 concentration affects water pH, which drops below 6 in the deeper parts of both lakes (Nicolosi, 2010). At lake surface, water pH varies seasonally be- tween 7 and 9 in both lakes (Fig. 4), depending on the in- tensity of algal photosynthesis (Dumontet et al., 2003; Ceccanti et al., 2007), which also affect oxygen content. In fact, during summer the high photosynthetic activity leads to O2 oversaturation (Dumontet et al., 2003). In LP, O2 concentration increases in summer from 11-12 mg L–1 at the surface to 23-25 mg L–1 at 5 m depth, dropping to zero below 20 m (Ceccanti et al., 2007; Spicciarelli and Mirauda, 2015). In LG, the seasonal pattern is similar, but the depth of water anoxia fluctuates between 5 m in sum- mer and 20 m in winter (Fig. 5), when a large portion of lake bottom, lying at 12 m depth, is oxygenated. A sea- sonal pattern also affects methane concentration in LG, but not in LP: Nicolosi (2010) reports that in the LP, the lowest methane concentration was measured in shallow Tab. 1. Reported Secchi disk depth (SD) and maximum macrophyte growing depth (MMGD) in Lago Grande and Lago Piccolo. Date Lago Grande Lago Piccolo Source SD (m) MMGD (m) SD (m) MMGD (m) July 1905 3.3 7 4.3 4 Forti and Trotter, 1908; Stegagno, 1908 Jun-Sep 1991 1.0-1.5 - 3-5 Marano and D’Aprile, 1991 Summer 1994 0.3 - “High” Schettler and Alberic, 2008 May 2001 0.5-1.0 - 3-3.5 Dumontet et al., 2003 February 2005 1.9 - - Ceccanti et al., 2007 August 2010 1.4 3.3 6.2 6.2 Azzella et al., 2010 September 2015 1.0-1.2 3.2-3.4 4.7-7 5-6 Spicciarelli and Mirauda, 2015 Fig. 5. Profiles of dissolved oxygen concentration in Lago Piccolo (left) and Lago Grande (right) in April, June, September, December 2005. Redrawn from Ceccanti et al., CNR-ISE 2007, with permission. No n- co mm er cia l u se on ly Literature record of Monticchio lakes 49 waters, increasing towards the lake bottom in the moni- molimnion, without a marked seasonal pattern. In LG, methane increases with depth and the highest concentra- tions in the entire hypolimnion have been measured dur- ing summer and autumn, while lower values were found in winter and spring. Water chemistry The most striking aspect of the ionic composition of LP water is the very high iron concentration (Fig. 6), reaching values higher than 100 mg L–1 in the anoxic deep water and causing lake meromixis (Ceccanti et al., 2007; Spicciarelli and Mirauda, 2015). High values of Mn (up to 2 mg L–1 in LP), Ba and Sr were measured in anoxic water in both lakes (Tab. 2). Such high iron concentration, coupled with the differ- ent thermal regime of the two lakes, dramatically affects the seasonal pattern of phosphorus concentration (Fig. 7). In LP, iron concentration in oxygenated water causes the formation of insoluble Fe3+ salts, which settle towards the deepest part of the lake, so that total P concentration in the upper 20 m of the lake never reaches 0.02 mg L–1. In LP anoxic bottom water, Fe3+ is reduced to Fe2+, and P is released, reaching a concentration of 0.9 mg L–1 in June 2005 (Ceccanti et al., 2007). In LG iron concentration is lower, and epilimnetic Fig. 6. Profiles of iron concentration in Lago Piccolo (left) and Lago Grande (right) in April, June, September, December 2005. Note the different scale. Redrawn from Ceccanti et al., CNR-ISE 2007, with permission. Fig. 7. Profiles of total P concentration in Lago Piccolo (left) and Lago Grande (right) in April, June, September, December 2005.Note the different scale. Redrawn from Ceccanti et al., CNR-ISE 2007, with permission. No n- co mm er cia l u se on ly R. Spicciarelli and A. Marchetto50 total P concertation reaches 0.06 mg L–1. During summer, in the anoxic hypolimnion total P concentration increases to values over 2 mg L–1 (Ceccanti et al., 2007; Spicciarelli and Mirauda, 2015). Winter overturn causes on one hand a small increase in bottom O2 concentration, causing a drop in total P concentration. On the other hand, water mixing leads to a P flux from deep water to surface water. During summer, total P concentration increases, suggest- ing the presence of local sources, such as the touristic ac- tivities on the lake shore. Total nitrogen profiles (Fig. 8) compare with the total phosphorus ones. Most inorganic N is in the form of nitrate in oxygenated water and of ammonium in anoxic water. Epilimnetic total N concentrations (0.5 to 3 mg L–1) are high enough to assure that N is not a limiting factor for algal growth (Teubner & Dokulil, 2002). Tab. 2. Range of metal concentration in Lago Grande and Lago Piccolo in 2005. Metal (µg/L) Lago Piccolo Lago Grande Min Max Min Max Al 5.25 55.9 7.8 182 As