Discussion 
In freshwater environments, biodiversity loss is dangerously menaced by several anthropogenic pressures acting on streams and river ecosystems both at local and global scales. This study is the first to examine possible temporal changes in hydro-morphology and aquatic biodiversity in an alpine protected area. Our study demonstrated that, overall, the hydro-morphology of the examined stretches within the Gran Paradiso area did not significantly vary over the last 15 years, as demonstrated by the indices derived from the RHS. HQA and HMS values indicated the presence of some modifications in the bank structure and the re-sectioning of the river course, and the strong lotic character which in some cases reduces the habitat diversification. The slight decrease in HQA or increase in HMS recorded in a few sites is mainly due to a decrease in the river flow and channel heterogeneity, as in the case of Leviona, or to bank reinforcements, as in Forzo, Roc, and Nampio. In some cases, HMS improved thanks to the removal of artificial structures such as small bridges or weirs. The stronger loticity highlighted by LRD Descriptor in 2020 confirms the highest percentage of erosive habitats detected during the last survey. Indeed, only 28% out of 194 monitored spot-checks showed erosive bank features in 2005 (such as local erosion, eroding cliff, or eroding bank), while 66% in 2020.
Water quality highlighted a small increase in the nitrate concentration in 2020 with a mean value of 0.760 mg l-1 that still corresponds to a first quality class according to the Italian legislation (DM 260/2010) and that is in line with nitrate concentrations in streams originating from ablating glacier (Brighenti et al., 2019). This increase in nitrates could be related to the ice melting as glaciers accumulated several atmospheric delivered chemical compounds and nitrate is easily eluted from melting glacier ice and snowpack (Saros et al., 2010). Indeed, the climatic data, although limited as temporal series, seem to indicate a trend towards an increase in maximum temperature: in recent years the summer temperature reach values close to 26 °C, and also in the winter months the values are always higher than in the 1999-2005 period. Also, the annual trend of the average monthly rainfall shows clearly visible differences, with higher values almost always in the years preceding the 2005 samplings. These data are in line with what has been highlighted by studies on climatic trends in the Alps, which are identified as an area particularly affected by climate change (Doretto et al., 2018; Gobiet et al., 2014).
Overall, our study shows a picture of few changes in local hydro-morphological impacts, accompanied by slight variations in water quality in a period affected by visible consequences of climate change,
How the stream benthic communities, and especially diatoms, respond to these changes is evidenced by the multivariate analysis of the community composition rather than by the traditional indices. The NMDS ordination technique performed on diatom community composition showed that the 2020 samples are nested in a wider ellipsoid including the 2005 data and resulted in significantly less heterogeneous. The species turnover between the two years mainly concerns co-generic species, mostly within the genera Navicula  and Nitzschia.  A. lineare , which showed a positive trend from 2005 to 2020, and A. minutissimum  have the highest SCBD scores: the two species however have different ecological preferences: A. lineare  is more sensitive to anthropic pressures, while A. minutissimum,  despite being very common in mountain streams, has a wide tolerance range towards both water pollution and physical alterations (Brighenti et al., 2019). It is important to notice that in 2005 species classified as “early alert” in the diatom Red List represented on average 6.6% of the total abundance, while only 3.1% in 2020. Of the 11 taxa belonging to this category, F. arcus  showed the most significant decrease from a mean abundance of 5.8% in 2005 to 2.2% in 2020. Another species deserving attention is Psammothidium subatomoides,  recorded in 8 samples out of 12 in 2005 and disappeared in 2020 inventories. The results obtained in this research confirm the need to preserve habitat conditions favoring the recovery of these species. When looking at species not yet classified in the Red List because of insufficient data, we focused on those species following a decreasing trend over the two sampling periods. Among them, Fragilaria delicatissima , detected in 2005 was not present in any of the 2020 inventories while Encyonopsis minuta  frequency showed a decrease. Being both very sensitive species, particular attention to their temporal and spatial distribution pattern should be paid. We found a general positive trend between the increasing habitat quality (HQA), diatom species richness and Shannon diversity, confirming that habitat heterogeneity is one of the main drivers for taxa richness (Bona et al. 2008). At the same time, this trend was negative (although not significant) considering habitat modification score, which means that poor communities are typical of the most impacted sites. Moreover, according to Ossyssek, Hofmann, Geist, & Raeder (2022), we found a good correlation between total diatom richness and endangered species.
Among the biomonitoring indices, a significant worsening of the ecological status was detected only by the TI index which on average indicated oligotrophy in 2005 and mesotrophy in 2020. The IPS remains in the class of oligotrophy and the ICM indicates a high ecological quality for both years. TI is the most sensitive to nutrient levels of natural origin (Rott et al., 1999) and it is reputed for its suitability in alpine rivers.
Macroinvertebrate communities showed no significant variations between the two campaigns. Orophilic and stenoxybiont organisms, strictly adapted to cold and harsh environments, characterized these communities on both occasions. It is interesting to notice that, regardless of the sampling year, macroinvertebrate biomonitoring indices showed a greater correlation with hydro-morphological alteration than those based on diatoms.
These differences in the response to anthropogenic stress confirm what has been reported by previous studies which compared the two indicators (Bona et al., 2008; Mykrä et al., 2012; Piano et al., 2019): diatoms respond to water quality and habitat features at different scales respect to macroinvertebrates, discouraging the use of one group as a substitute for the other. Their different responses are undoubtedly linked to bio-ecological factors, such as their different trophic role, and the difference in size and length of the life cycle. Metrics related to diversity are certainly influenced by the taxonomic level and the qualitative or quantitative nature of the data collected as well. Indeed, indices currently used for macroinvertebrates require the genus as the finest taxonomic level and data collected in 2005 were expressed as presence/absence. For diatoms, the metrics used are based on the taxonomic level of species and their relative abundance and this could have contributed to their more complex response highlighted by the multivariate analysis.
Our study highlights the role of a large, protected area in preserving freshwater habitats and biological communities from human impacts. The ß- diversity partitioning identified sites to prioritize in protection efforts, confirming its role in conservation planning (Ruhı´, Datry, & Sabo, 2017).
Moreover, this research demonstrates that the creation of databases relating to the ecological quality of rivers is crucial for tracking the environmental evolution of lotic systems, creating historical series of considerable management interest, and supporting decision-making strategies also in rural or mountainous areas. In this context, the presence of a natural park is essential for its role as a promoter of research and data collection.
Indeed, the conservation of aquatic ecosystems is one of the main objectives of protected areas and requires harmonization actions among protected areas managers and stakeholders, who should be driven toward biodiversity protection. In protected areas, water use for human purposes (such as recreational or energetical) is generally minimal, and catchment integrity is mostly maintained, so as spatial and temporal river connectivity and habitat heterogeneity. All these features significantly contribute to the maintenance of a high functional richness of biological communities, as demonstrated by recent research (Falasco, Piano, Doretto, Fenoglio, & Bona, 2018; Falasco, Doretto, Fenoglio, Piano, & Bona 2020; Falasco, Bona, Risso, & Piano, 2021; B-Béres, Stenger-Kovács, Török, & Török-Krasznai, 2021). It has been recognized that high functional richness increases community resilience against extreme events, which is expected to significantly increase under a global climate change scenario (Mason, Mouillot, Lee, & Wilson, 2005; Doretto et al., 2020; B-Béres et al., 2021). However, not all protected areas can play this role with the same effectiveness with rivers. In fact, the GPNP covers the entire headwater and therefore there are no impacts located upstream.
In conclusion, this study once again underlines the importance of maintaining and, if necessary, enlarging the alpine protected areas facing an era of rapid climate change (Finlayson & Pittock, 2018), as they play a fundamental role in the preservation of the water resource and the ecosystems that guarantee its quality.