Climate‐induced plasticity in leaf traits of riparian plants

Leaf litter inputs from riparian vegetation and its decomposition play a key role in energy and nutrient transfer in many stream ecosystems. Instream leaf litter decomposition is driven by both leaf traits and environmental conditions. Therefore, understanding and predicting leaf trait variation under current environmental changes and their putative interactive effects on stream food webs is a critical challenge. Most studies have focussed on the assumed higher interspecific leaf trait variability, with little research addressing an intraspecific perspective.


| INTRODUC TI ON
The warming of the Earth system is unequivocal (IPCC, 2021).
Globally, precipitation is also predicted to increase in the long term (Hewitson et al., 2015). However, forecasts in the Mediterranean basin point to a precipitation decrease of around 34%, along with a temperature increase of ca. 5°C for the period 2000-2099 (Harris et al., 2013). As a result, this region will face a climate much drier and hotter than at present, especially during warm seasons (Giorgi & Lionello, 2008), with direct effects on hydrologic regimes (Nohara et al., 2006;Vicente-Serrano et al., 2014) and soil moisture (Manabe et al., 2004). These changes may alter the functioning and structure of plant communities (e.g. Carnicer et al., 2011;Trivedi et al., 2008;Vicente-Serrano et al., 2012).
Small streams flowing through forested areas can be especially susceptible to climate change-induced alterations in plant communities, owing to their high dependence on organic matter inputs from the riparian vegetation, i.e. leaf litter (Wallace et al., 2015). Instream decomposition of leaf litter is a crucial ecosystem process, involving the cycling of nutrients and fuelling stream secondary production (Marks, 2019). The rate at which leaf litter decomposes and is incorporated into food webs highly depends on its quality, which fundamentally depends on after-life persistent traits (Graça & Cressa, 2010;Graça et al., 2001;Zhang et al., 2019). Thus, ecosystem functioning can be significantly altered if leaf litter inputs to streams experience physical and chemical changes (e.g. Casas et al., 2013;del Campo et al., 2021;López-Rojo et al., 2019). These changes can be interspecific, e.g. resulting from the forecasted substitution of deciduous by evergreen species (Kominoski et al., 2013;Salinas et al., 2018) and/or the decline of key plant species populations (e.g. alder; Alonso et al., 2021;Rubio-Ríos et al., 2021). Moreover, given that leaf traits are highly responsive to environmental changes (Heilmeier, 2019;Soudzilovskaia et al., 2013), intraspecific changes may also occur, e.g. due to genetic variability (Crutsinger et al., 2014;LeRoy et al., 2012) or phenotypic plasticity (Graça & Poquet, 2014;Henn et al., 2018;Jung et al., 2014).
Such relationship between leaf traits and the environment has been a recurrent theme of the study (e.g. Ordoñez et al., 2009;Read et al., 2014;Reich & Oleksyn, 2004). However, although recent results indicate that intraspecific variation may represent up to ca. 30% of total functional trait variability in plant communities (Albert et al., 2010;Siefert et al., 2015), most studies have focussed on the often assumed higher interspecific variability of many leaf traits (e.g. Hulshof & Swenson, 2010;Wright et al., 2004).
High rates of plasticity in leaf traits are expected in species distributed across ample environmental gradients (Cordell et al., 1998;Fajardo & Piper, 2011;Umaña & Swenson, 2019), as increases in niche breadth allow plants to respond to variation in climatic and other environmental conditions (Henn et al., 2018), whereas nearby individuals may share biotic and abiotic pressures and have close genetic relationships. Warming and reduced rainfall, i.e. increasing aridity, are usually reported to promote the production of thicker and smaller leaves (Wright et al., 2004)-in order to improve their water use efficiency and to increase their leaf life span-with low nutrient concentrations (Reich & Oleksyn, 2004). Such plasticity in important traits can, in turn, affect the palatability and decomposability of leaves, i.e. their acceptability and easiness to be consumed, along environmental gradients (Boyero et al., 2017;Graça & Poquet, 2014;Lecerf & Chauvet, 2008;LeRoy et al., 2007). Understanding how individual species traits, or their syndromes, are modulated by climatic or other environmental characteristics could allow us to refine predictions of potential effects on stream ecosystem functioning, both in green (based on primary production) and brown (based on detritus) food webs, in the face of climate change (Kominoski et al., 2021).
Here, using a 'space-for-time' (SFT) substitution approach (Blois et al., 2013;Pickett, 1989), we investigated how climate change might affect leaf quality, focusing on after-life traits affecting leaf decomposition. The SFT substitution approach is a useful tool to anticipate changes taking advantage of natural gradients (Fukami & Wardle, 2005); in the present study, a natural aridity gradient represents the forecasted aridification of the Mediterranean basin (Seager et al., 2014). We assessed plasticity in leaf traits of four common riparian species, with contrasting functional traits, in permanent low-order streams [Alnus glutinosa (L.) Gaertn., Salix atrocinerea Brot., Rubus ulmifolius Schott and Nerium oleander L.], extrapolating their possible variation in the forecasted climatic scenarios from that observed across a wide environmental gradient studied within a relatively small region. Using the same species along many areas differing in environmental conditions allowed us to control for species-specific traits, but not to assess the amount of trait variability due to genetic variability.
Given the high responsiveness of leaves to climate changes (Heilmeier, 2019;Soudzilovskaia et al., 2013) and the high water and nutrient availability in riparian soils of permanent streams (Naiman & Decamps, 1997), we hypothesize that (1) climate will exert a higher influence on leaf trait plasticity of the studied species compared to soil variables. We (2) expect a general trend of decreasing leaf quality-i.e. lower nutrient concentration, higher toughness-with the forecasted aridification (increasing temperature and decreasing precipitation) (Reich & Oleksyn, 2004). However, we also expect that the strength of the effects will vary among different species, as they belong to different functional groups (i.e. C allocation and/or N-fixing) and therefore have low similarity in their leaf traits (Salinas et al., 2018). Thus, we also hypothesize that (3) trait plasticity will be relatively low compared to interspecific variation.

| Area of study and selected plant species
Our study was conducted during summer 2013 in the riverbanks of 34 headwater streams with permanent flows distributed across nine natural protected areas (considered as pristine) located in Andalusia (south of the Iberian Peninsula), covering ca. 88,000 km 2 . These locations represent a wide climatic gradient within the context of a Mediterranean-type climate and possess a considerable lithological and topographical heterogeneity (Figure 1). The present (mean annual temperature range 10.8-17.4°C; mean annual precipitation range 261-845 mm; Table S1) and the projected climatic gradient studied (by the end of the 21st century) covers from arid to humid conditions according to the Emberger's bioclimatic coefficient (Table 1, Figure 2). This embraces the forecasted aridification, i.e. warming (mean temperature rise of 2-4°C) and reduction of precipitation (mean precipitation decrease of 10-40%), for the Mediterranean region (Seager et al., 2014), as a consequence of climate change towards the year 2100 (reviewed by Giorgi & Lionello, 2008).
We selected four abundant riparian plant species which represent different functional groups featuring different characteristics, including two deciduous riparian trees: black alder-Alnus glutinosa (an N-fixer), and grey willow, Salix atrocinerea; one semi-deciduous shrub: blackberry, Rubus ulmifolius; and one evergreen shrub: oleander, Nerium oleander, also known as laurel rose. Leaves of these species collected (June-July 2013) from each sampling sites were present (Table 1) from robust, well-grown and totally unshaded plants distanced from the stream by a maximum of 6 m. Those leaves directly exposed to sun light and without herbivory or pathogen symptoms were selected (Cornelissen et al., 2003). In each stream and for each species, we collected 102 leaves from six individuals (17 leaves per individual) randomly distributed on both stream sides along a 100 m stream reach. Leaves were air-dried at room temperature (20-23°C) for one week and stored in darkness in paper bags until processed. At each stream, the cover of each species was estimated using the Domin-Krajina scale of cover and abundance (Kent & Coker, 1992) in six plots (36 m 2 each) randomly distributed in both stream sides-three plots per side arranged from the edge of the wetted channel-along a 100 m stream reach (Salinas et al., 2018).

| Environmental variables
Thirty-two environmental variables (altitude, 20 climatic and 11 edaphic; Table 1 and S1) were selected as potential predictors of leaf trait plasticity. Altitude was obtained in situ using a portable GPS. scenario, CCM3) for the year 2100 (Govindasamy et al., 2003) and subsequently downscaled and matched to the WorldClim estimates of current climate at a resolution of 2.5 minutes (i.e. ~4.5 × 4.5 km).
From these variables, the Emberger's bioclimatic coefficient (Q2) for each site was calculated following Condés and García-Robredo  (Table S1) were measured as in Gil et al. (2004).

| Leaf traits
We measured nine leaf traits that often correlate with leaf litter decomposition rate (see Graça et al., 2015;Tonin et al., 2021)  Concentrations of C and N (% dry mass, DM) of leaves were determined using a mass spectrometer (EA-Thermo DELTA V Advantage, Fisher Scientific ® ) following standard procedures (Flindt et al., 2020).
The concentration of P (% DM) was measured spectrophotometrically after autoclave-assisted extraction (APHA 1998, Flindt et al., 2020. Concentrations of Ca and Mg (% DM) were determined by inductively coupled plasma mass spectrometry (ICP-MS, Perkin Elmer DRC II). Condensed tannins (mg Catechin Hydrate Equivalent per g of DM) were measured by the acid butanol assay (Gessner & Steiner, 2020). Concentration of lignin (% DM) was estimated gravimetrically using the acid detergent method of Goering and Van Soest (1970).

| Data analysis
To elucidate the relationships between species cover and environmental variables, we ran a Canonical Correspondence Analysis   Mehmood et al., 2012). Those variables with VIP ≳1 were considered relevant (Andersen & Bro, 2010). Spearman rank correlation analyses were used to equalize the size of the two matrices of environmental variables removing those variables with high collinearity within those with higher VIP values ( Figure S2, Tables S3 and S4).
A second PLS regression was performed for each species using the selected variables, and the influence of each group of environmental variables (climate and soil) and their combination (climate + soil) on leaf plasticity was assessed using the goodness of prediction (Q 2 ) and the goodness of fit (R 2 (Y)) of models. A model was considered significant when Q 2 > 0.097 (Friden et al., 1994).

TA B L E 2
Summary of univariate dependent variable PLS models fitted to the first two principal components of PCA (PC1 and PC2), summarizing leaf trait plasticity for each species, using three matrices (C, S and C+S) of selected (in preliminary PLS regressions) environmental variables as predictors

| Environmental variables and species distribution
Overall, the four species covered a large gradient of climatic conditions from semi-arid to humid bioclimatic types according to the  and Salix with the most constricted distribution (44.6% of CCA1 and 71.7% of CCA2) ( Table 1, Tables S1 and S2; Figure 3).

| Interspecific variation and species plasticity of leaf traits
Species differed significantly in all leaf traits measured (one-way ANOVAs, all p < .0001) (Figure 4, Table S6). Alnus showed the lowest toughness and the highest N concentration, and consequently the lowest C:N ratio, being for these traits antithetical to Nerium, which in turn showed the highest Ca concentration and C:P ratio. Overall, interspecific variation was higher than trait plasticity ( Figure 5a). Rubus, the most widely distributed species, showed higher trait plasticity on PC 1, occupying 54% of this leaf quality gradient while other species ranged between 23% and 38%. However, the two species with more restricted distribution, Alnus and Salix, showed the highest trait plasticity on PC 2, occupying 66% and 51% of this leaf quality gradient, respectively, compared to the more widely distributed Rubus and Nerium (both 40%) (Figure 5a).
Regarding individual traits, variance partitioning analyses indicated, overall, higher interspecific variation than species plasticity in leaf traits ( Figure S3). The highest interspecific variation (>80%) occurred in traits considered major determinants of litter decomposability-palatability-toughness, lignin, N and C:N-as expected dealing with species across different plant functional types.
However, trait plasticity was higher than interspecific variation for P, Ca, Mg and C:P (ranging between 55% and 71%) and noticeably high for tannins ( Figure S3).

| Relative importance of climate and soil factors, and best climatic predictors of leaf trait plasticity
Univariate dependent variable PLS models indicated that leaf trait plasticity (PC 1) of the four species responded significantly and predominantly to climatic variables (Table 2). Adding soil factors to climate increased noticeably the goodness of prediction in Nerium, but produced a highly complex model with six latent variables. Models predicting leaf trait plasticity associated to PC 2 were only significant for Nerium and Salix, but especially for the latter, in which the set of soil variables significantly predicted a high proportion of variance of leaf trait plasticity, but the model including just the set of climate variables was still significant (Table 2).
Overall, climatic predictors with the highest influence (VIP close or >1) on leaf trait plasticity associated to PC 1 (Table 3) varied among species, although most notable differences arose between broad functional groups. Mean temperature of the wettest quarter (late winter-early spring) was an important predictor with negative effects on leaf quality for deciduous/semi-deciduous species.
Conversely, maximum annual temperature was the main predictor with high positive effect on leaf quality for the evergreen Nerium.
Temperature annual range was an important predictor of leaf quality (PC 1) for Nerium and Alnus, although with contrasting sign (negative and positive, respectively), highlighting the opposite response that species belonging to different plant functional types may have the same climatic variable. Moreover, precipitation variables (Table 3) did not have substantial effects on the evergreen Nerium, but were important predictors of leaf quality (PC 1) for deciduous/semideciduous species, with notable positive effects on Salix and Rubus, but slightly negative on Alnus. Leaf quality of Salix associated to PC 2 was primarily predicted by temperature annual range (positive effect) and winter temperature (negative effect), with precipitation variables (Table 3) being other important predictors with positive effects on leaf quality. Over this dimension, soil EC and P (with negative effects) and soil CaCO 3 (with positive effects) were important predictors on leaf quality of Salix.

| Forecasted intraspecific changes in leaf quality induced by climate change
Our modelling projections showed that the four plant species would respond differently to the forecasted scenario of aridification by the year 2100 (2 × CO 2 climate change scenario) in the studied region, although with remarkable congruence within broad functional groups in terms of response direction (Figure 6; Figure S4). For Alnus and Salix (PC 1), we observed weak evidence of overall variation in leaf quality (t = 1.523, p = .154; t = −2.071, p = .065, respectively; Hedge's g = 0.232 and −0.295, respectively; Figure 6). Salix (PC 2;

| DISCUSS ION
Functional trait-based approaches are potentially useful to understand how species respond to environmental changes (Soudzilovskaia et al., 2013;Zhang et al., 2020) and, therefore, are important for an ecologically sensitive management of ecosystems.
Here, we assessed how climate change might affect leaf quality of different riparian woody species from an intraspecific perspective, which has been much disregarded based on the general assumption that intraspecific variation accounts only for an irrelevant portion of total trait variability (Garnier et al., 2001). Overall, in support of our first hypothesis, but contrary to previous studies (Graça & Poquet, 2014;Ordoñez et al., 2009), climate showed larger influence than soil explaining most leaf trait plasticity. Our second hypothesis of decreasing intraspecific leaf quality-linked to determinant traits of palatability and decomposability-with increasing aridity was partially supported, given that increasing temperature had negative effects on leaf quality of deciduous and semi-deciduous species, but not on the evergreen Nerium, which displayed the opposite response.
These results suggest potential effects on stream ecosystem functioning Martínez et al., 2013), but with inverse sign depending on the identity of dominant species in the riparian vegetation. Moreover, in support of our third hypothesis, we generally observed higher variation among species than plasticity within F I G U R E 4 Box-and-whisker plots for selected leaf trait variables of the four plant species studied: nitrogen (N), phosphorus (P), calcium (Ca) and magnesium (Mg) concentrations (% DM), molar elemental ratios (C:N and C:P), lignin concentrations (% DM), condensed tannins concentrations (mg Catechin Hydrate Equivalent g DM −1 ) and toughness (g) of each plant species. Box represents median and 25th and 75th percentile levels, crosses are the mean, whiskers are the range, and dots are replicates. Different letters indicate significant differences (p < .05) among plant species, on the basis of linear models followed by pairwise multiple comparisons (Tukey test) species, except for a few traits (e.g. P, Ca and Mg concentrations and C:P ratio) that exhibited remarkable leaf trait plasticity (Albert et al., 2010;Fajardo & Piper, 2011). Nonetheless, ranges of trait plasticity found here for some traits (e.g. %N, %P and %lignin) are similar, or higher, than those reported before for other species (e.g. Lecerf & Chauvet, 2008;LeRoy et al., 2007;Oliveira et al., 2021).

| Relative influence of climate and soil factors
Over the environmental gradient studied, climate exhibited an overall higher influence than soil on most species' leaf trait plasticity, although soil was the strongest predictor in some cases (e.g. Salix and Nerium PC 2). We presumed higher responsiveness of leaf traits to climate than soil in species with distributions highly constrained by soil conditions. This appears to be the case for the acidophilic Alnus (Miles, 1985), the species with the highest control of climate on its leaf trait plasticity. The fact that Alnus is an N-fixer may have further contributed to make this species less sensitive to soil nutrients. However, other species with less restricted soil-related distributions, such as Rubus-spread out across almost the entire study area-or Nerium, also showed a prominent role of climate influence on leaf trait plasticity. Similar patterns have been observed when assessing the abundance of plant functional types in the same region and across environmental gradients (Salinas et al., 2018). This lower predictive role of soil variables may stem from the high dynamics of alluvial soils and their permanent water availability, which would tend to homogenize conditions-i.e. nutrient availability-among sites (Naiman & Decamps, 1997). Yet our results are counter to other findings recorded at much larger spatial scales that observed substantial importance of soil predictors explaining intraspecific changes in leaf traits (Graça & Poquet, 2014;Ordoñez et al., 2009).
This suggests that other factors not considered here, such as the great topographic variability present in our spatial gradient, or genotype differences, might be overriding soil effects.

| Main climatic predictors of leaf trait plasticity
Among climatic the factors, temperature exhibited much clearer patterns than precipitation on the main dimension of leaf trait plasticity (PC 1). This is to be expected in riparian belts of permanent streams where soil moisture tends to be relatively high and constant in the absence of extreme drought events (Moore et al., 2016), preventing major water stress in plants and its consequences on leaf characteristics (e.g. LeRoy et al., 2014). However, climate-driven changes in streamflow may worsen the effects of aridification on such ecosystems (Perry et al., 2012).
Despite clear differentiation in distribution extent among species, we detected a common negative relationship between temperature and leaf quality in the deciduous and semi-deciduous species. On the contrary, this relationship was positive for the evergreen Nerium.
Overall, nutrient concentrations (N, P, Ca and Mg) decreased, but tannin and/or lignin concentrations, and/or toughness increased with increasing temperature for deciduous/semi-deciduous species, whereas Nerium roughly exhibited the opposite pattern. Thus, within the frame of the leaf economic spectrum (Reich et al., 1997;Wright et al., 2004), the above seems to reveal antithetical syndromes of leaf traits between functional groups in response to temperature, in which the intraspecific intercorrelated leaf traits along our quality gradient represent physiological and structural trade-offs (Boyero et al., 2017;Onoda et al., 2017).
Decreasing leaf N and/or P concentrations with increasing temperature has been reported before in woody deciduous species (Chen et al., 2011;Kudo et al., 2001;Sun et al., 2015). This may be explained by an increase of the catalytic capacity of photosynthetic enzymes at higher temperatures, requiring lower enzyme amounts (e.g. lower N concentration) to maintain photosynthetic rates (i.e. the photosynthetic rate is achieved with lower amounts of such enzyme; Scafaro et al., 2017). Alternatively, or additionally, higher temperature is often associated with increasing length of the growing season in deciduous species, which in turn promotes long leaf life span.
Reduction of nutrients and strengthening of leaf traits to confer resistance (e.g. increasing toughness) have been reported in evergreen Quercus species in response to decreasing winter temperatures. This is interpreted as a higher cost for evergreens at cooler sites compared with deciduous trees (González-Zurdo et al., 2016).
However, this finding is not totally consistent with our results for Nerium as winter temperatures did not exhibit any effect on its leaf quality. We observed the strongest positive effect on leaf quality of Nerium from maximum temperature, but a more negative effect from annual temperature range. This suggests that Nerium develops more nutrient-rich and softer leaves in its optimum distributional range (areas with mild winters and maritime influence), with negligible effects from harsh low-winter temperatures, which are infrequent in its area of distribution. Nevertheless, we cannot rule out the possibility that our results are species-specific, and projection of such results to the entire functional group needs to be confirmed with the study of further evergreen species.
A substantial amount of leaf trait plasticity (25%) in Salix (PC 2)-positively related to leaf N and lignin, and negatively to Mg concentrations-was significantly explained by climatic conditions, but much more by soil variables. The strong positive association of N and lignin on PC 2 suggest that this N fraction is structural, possibly lignin-bound N, therefore not readily available to decomposers and detritivores (Berendse et al., 1987). Thus, PC 2 represents a structural reinforcement of Salix leaves positively related with temperature annual range and negatively with winter temperature, but also, and mostly, negatively with soil P. A structural reinforcement of leaves (increasing leaf mass per area and lignin concentration) with decreasing soil fertility has been documented elsewhere (e.g. Diehl et al., 2008).
The trait plasticity observed in this study can arise from responses to environmental conditions, but also from genetic variability. Genotypes, although largely influenced and selected by local environments, represent an important source of trait variability unaccounted for here. Genetic variability has been exhibited to strongly influence litter quality and, consequently, associated ecosystem processes (e.g. litter decomposition) and communities (Crutsinger et al., 2014;LeRoy et al., 2006LeRoy et al., , 2007LeRoy et al., , 2012. Given that leaf traits differ in their heritability, for example, tannins appear to be highly heritable  (Jackrel & Morton, 2018;Jackrel et al., 2016;Lecerf & Chauvet, 2008;LeRoy et al., 2007;Oliveira et al., 2021). Here, we assessed the plasticity of selected traits of green leaves of riparian plants aimed at forecasting potential consequences of climate change on stream ecosystems highly dependent on these resources (i.e. forest streams; Wallace et al., 2015). Although inputs of leaves to streams are mainly in the form of leaf litter, it has been reported that some traits of green leaves tend to persist after senescence and control rates of litter decomposition (Cornelissen et al., 1999;Cornwell et al., 2008). Therefore, if nutrient resorption efficiency remains fundamentally invariable across climatic conditions (Norby et al., 2000, Aerts et al., 2007, but see Yuan & Chen, 2009b), understanding how green leaves respond to climate change may allow us to anticipate effects of leaf quality changes on stream ecosystem functioning. In support of this idea, a recent study suggests that traits of green leaves can be used to accurately predict decomposition rates (Rosenfield et al., 2020). However, as others have pointed out that traits of litter can differ from those of fresh leaves (Hättenschwiler et al., 2008;Hättenschwiler & Vitousek, 2000;Horner et al., 1987;Yuan & Chen, 2009a), the potential effects on headwater stream functioning exposed here should be interpreted with caution.
Litter decomposition is often reported to be enhanced by its high N and P concentrations (García-Palacios, McKie, et al., 2016;MacKenzie et al., 2013). Elevated litter concentrations of Ca and Mg-reported to be important for fungal decomposers (Jenkins & Suberkropp, 1995) and macroinvertebrates (Makkonen et al., 2012;National Research Council, 2005)-can also accelerate decomposition (Santonja et al., 2019). Moreover, tannins (Coq et al., 2010;Irons et al., 1988), lignin (Ferreira et al., 2016;Ramos et al., 2021;Schindler & Gessner, 2009) and toughness Li et al., 2009) primarily tend to reduce litter consumption by detritivores. Our results point to a general decrease in leaf quality as a response to aridification in the three deciduous/semi-deciduous species. This decrease was generally related to a reduction in leaf N and P, but also Ca and Mg, versus an increase in tannins or lignin, and leaf toughness.
In particular, changes in leaf quality of the deciduous N-fixer Alnus could have major consequences given the key role of this species on stream ecosystem processes (Alonso et al., 2021;Rubio-Ríos et al., 2021). We reported here for Alnus ranges of %N, %P and %lignin variation similar to those reported at the European continental scale (Lecerf & Chauvet, 2008), and 53% of its species leaf trait plasticity was remarkably explained by climatic variables, yet our forecasted decrease in leaf quality was relatively low (11%) and not statistically significant, compared to other species. Nonetheless, apparent subtle changes in litter traits might result in major effects in consumer fitness .
Furthermore, this projected minor decrease in leaf quality adds to the decline of populations of this key species through Europe due to a disease caused by the pathogen Phytophthora alni (Bjelke et al., 2016), which also has been recently reported to alter the nutritional quality of leaf litter . Both factors are likely to trigger significant alterations to the functioning of forested streams (Alonso et al., 2021). Moreover, if a general decrease in leaf quality occurs in other deciduous species, as those forecasted here for Salix and Rubus, the negative influences on stream food webs will increase.
Thus, our results indicated that decreases of leaf quality of individual deciduous species may occur in a relatively short term (via phenotypic plasticity; Nicotra et al., 2010; but see Valladares et al., 2007), which in the long term will add to the forecasted dieback of deciduous woody species in riparian corridors of temperate climate zones (Kominoski et al., 2013;Salinas et al., 2018). Both riparian changes have the potential to significantly impair instream ecosystem processes, particularly in mountain streams presently dominated by deciduous vegetation , more than in lowland streams where deciduous species actually represent a minor component of the riparian belt. feedback that significantly improved this manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/ddi.13493.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data openly available in a public repository https://doi.org/10.5061/ dryad.bzkh1 899h.