The nutrients entering the forest floor via rainfall is altered by the canopy and the litter layer. When rainfall enters a forest canopy, a portion of gross rainfall ( GR ) reaches the forest floor by dripping from vegetation or by passing directly through tree canopies as throughfall ( TF ). The remaining GR either reaches the forest floor by flowing along stems as stemflow ( SF ) or evaporates back to the atmosphere ( Hanchi & Rapp, 1997 ; Sadeghi et al. , 2016 ). The chemical compositions of TF and SF changes after contacting canopy elements. In addition to the nutrients from wet deposition, TF and SF can absorb nutrients present in the canopy. The nutrients may transfer to the available nutrient pool in the soil thereby affecting forest soil fertility ( Levia & Frost, 2003 ) and nutrient dynamics ( Parker, 1983 ). Upon entering the O layer, the chemical composition of the rainwater is further altered ( Levia & Frost, 2003 ).

Forest canopies alter the chemical composition of precipitation due to an interaction between precipita­tion and the crown of the trees. By restoring degra­ded forests with a nonnative species, the concentration of elements, pH, and electrical conductivity (EC) in TF are altered ( Eaton et al. , 1973 ). In addition, the nutrient concentration can be affected by other factors such as stand density, canopy structure, rainfall intensity, rainfall continuity, rain angle, crown size, branch shape, branch angle, bark, and the nutrient content of atmospheric rainfall ( Cattan et al. , 2009 ). For example, coniferous forests at the same site and under the same climatic conditions intercept more atmospheric pollutants than deciduous forests annua­lly ( De Schrijver et al. , 2007 ).

Forest nutrient cycles are linked to the hydrological cycle because water acts as the main transporting agent and solvent for nutrients. Rainfall is a considerable source of nutrients for forest ecosystems and plays an important role in the transfer of material from the canopy to the litter and mineral soil. Three processes are generally linked to the change in elemental concentrations of precipitation ( Parker, 1983 ): (1) accumulation of atmospheric suspended solids on the surface leaves and branches, (2) secretion of plant tissues to the outer surface of leaves and branches, and (3) absorption of chemical nutrients by the foliage (plant tissues). Plant leaves have retention and adsorption capacities for atmospheric particulate pollutants because of their unique surface characteristics and leaf distribution ( Schaubroeck et al. , 2014 ; Fan et al. , 2015 ). The leaf surfaces collect nutrients because of evaporation from the leaf surface, the absorption of particulate matter by leaves and by the accumulation the plant secretions on the leaf surface. The quantities of these elements differ depending on the type and characteristics of plant species, topographic and cli­matic conditions ( Carlyle-Moses et al. , 2004 ). Some substances enter the leaf via passive processes driven by concentration gradients ( Fernández & Eichert, 2009 ). The remainder remains on the leaf surface and can alter the chemical composition of TF ( Adriaenssens et al. , 2012 ). Hence, canopies are both a sink and a source of nutrients ( Lovett & Lindberg, 1984 ). The changes of elements concentration depend on the type of forest (conifers or hardwoods), forest structure and ecological and climatic factors ( Iida et al. , 2005 ; Herbst et al. , 2007 ).

Chiwa et al. (2004) studied the chemical elements of the TF in a Picea sitchensis plantation in six different forest habitats with intense air pollution in China. They concluded that EC, Ca ²⁺ , K ⁺ , Mg ²⁺ , and Zn in the TF increased after passing through the canopy. Shen et al. (2013) also investigated the concentration of chemical elements and pH of TF under the canopy of planted stands of Acacia mangium and Dimocarpus longan in China and noted that the amount of pH in both stands were more than in GR . Abbasian et al. (2015) also showed that the concentrations of some elements increased after passing through the canopy in a Picea abies plantation and a Fagus orientalis natural stand.

In addition, the O layer interacts with rainwater and can provide cations to deeper soil layers horizons ( Eaton et al. , 1973 , Bernhard-Reversat, 1975 ). The amount of nutrients exiting the O layer strongly depends on the quantity of TF ( Ashagrie & Zech, 2010 ). The quantity of elements in the O layer in Fagus orientalis and Picea abies stands differs from elements exiting the O layer because of differences in the quality and quantity of net rainfall reaching the forest floor ( Hojjati et al. , 2009 ).

Change in the type of tree species in a geographic area creates significant changes in the composition of water entering the forest soil via precipitation ( Llorens & Domingo, 2007 ). Anatomy, morphology, and physiology in different species may also play a role in TF chemistry. Increased concentrations of plant nutrients in TF depends on canopy structure because the accumulation of particles, dusts, and gaseous compounds are generally higher in evergreen, coniferous canopies, than in deciduous species ( Draaijers et al. , 1992 ; De Schrijver et al. , 2004 ). Robson et al. (1994) suggested that temporal and spatial variability in TF chemistry between forest canopies is generally attributed to non-uniformity of canopy density in different species and to differences in the efficiency of different canopy structures for filtration dry deposition. Bhat et al. (2011) stated that TF quality and thus the amount of plant nutrients that reach the forest floor by TF depend on composition of tree species.

The Caspian forests of northern Iran were histori­cally comprised of broadleaved deciduous forests that covered an area of 1.8 million ha, contained 15% of the total forests of Iran and represented 1.1% of the country's area ( Abbasian et al. , 2015 ). It is a green belt stretching over the northern slope of the Alborz mountain ranges and covers the southern coast of the Caspian Sea ( Sagheb Talebi et al. , 2014 ). The forests began to degrade due to overexploitation of wood and livestock overgrazing in the past few decades. Since the 1960s, the Forest, Range, and Watershed Management Organization (FRWO) of Iran established restoration projects in an effort to restore the deciduous forests of northern Iran and to conserve water and soil ( Abbasian et al. , 2015 ). Degraded forests have been restored using native and indigenous species in northern Iran. Plantations of indigenous species are regarded as a viable management strategy for rehabilitation of na­tive tree communities ( Chapman & Chapman, 1996 ). However, many of the Caspian forests in northern Iran have been replaced with P. abies plantations. The increase in nonindigenous species may alter ecological process in these regions ( Abbasian et al. , 2015 ). When native deciduous forests are replaced with non-native coniferous species, the soil fertility and nutrient cycling can be significantly impacted. Planting nonnative tree species can alter water and nutrient cycling (Chiwa et al. , 2004 ; Shen et al. , 2013 ; Abbasian et al. , 2015 ). The objectives of this research were to compare and contrast how replacing a natural oriental beech forest ( Fagus orientalis Lipsky) in the Caspian region with a coniferous forest, i.e., Norway spruce ( Picea abies ), impacts nutrient concentrations (NO ₃ ^- , NH ₄ ⁺ , PO ₄ ^3- , Ca ²⁺ , K ⁺ , Mg ²⁺ , and Na ⁺ in mg L ^-1 ), pH, and EC of TF , SF , and the O layer.

In general, the pH of TF declines relative to GR . As with our research (Table 2), Adriaenssens et al. (2012) and Douglas et al. (1988) reported significantly lower pH in TF relative to GR in a P. abies and an Abies balsamifera stand, respectively. Unlike past research, there was no significant difference between pH in TF of beech and spruce stands in our study. In contrast, Hongve et al. (2000) reported that acidity of TF in conifer forests located in Norway was higher than that of a broadleaf forest.

Similar to past research, EC in GR was lower than that of TF, SF and O layer in both forests. Previous studies broadly report that the EC increases because of interactions with the canopy ( Chiwa et al. , 2004 ; Polkowska et al. , 2005 ; Wang et al. , 2006 ; Hermann et al. , 2006 ). For example, Polkowska et al. (2005) indicated that the EC of TF was higher than that of GR by about 0.03 to 0.05 dS m ^-1 . The increase in EC can be attributed to accumulation of charged dust particles and ions, such as Na ⁺ ( Chiwa et al. , 2004 ). The elements may be present on the foliage/stems of the trees as a result of dry deposition from tissues tree secretions ( Chiwa et al. , 2004 ).

After passing through the canopy, the concentration of Ca ²⁺ , K ⁺ , Mg ²⁺ and Na ⁺ in TF were significantly greater than GR (Fig. 3). The increased concentrations, e.g. Ca ²⁺ and K ⁺ , were generally higher in the spruce stand relative to the beech forest. Hojjati et al. (2009) stated that throughfall fluxes of most of the elements were considerably higher under the canopy of spruce compared with beech. The greater concentrations of Ca ²⁺ and K ⁺ might be attributed to the greater leaf area index ( LAI ) of the spruce forest ( De Schrijver et al. , 2007 ). The increase in surface area can result in a higher rate of leaching of these cations from the needles ( De Schrijver et al. , 2007 ; Tukey, 1970 ). In addition to the higher LAI , higher filtration capacity of spruce canopy and higher foliage longevity compared with beech are the main reasons for higher element fluxes in TF under spruce ( Hojjati et al. , 2009 ). In both species, Ca ²⁺ ions were likely washed from the crown of the trees. Past researchers report that the concentration of Ca ²⁺ in TF can increase by 5-8 times relative to GR ( Dezzeo & Chàcon, 2006 ). Researchers also reported that the K ⁺ cation is easily removed by precipitation, thereby increasing concentrations in TF and SF relative to GR ( Parker, 1983 ; Edmonds et al. , 1991 ). Adriaenssens et al. (2012) reported that the concentrations of Ca ²⁺ and K ⁺ in European beech and spruce forests were higher than that those of GR . Hojjati et al. (2009) stated that canopy leaching is the main reason for increasing Ca ²⁺ (50%), Mg ²⁺ (60%), and K ⁺ (90%) in TF of beech and P. abies stands.

Similar to our study, Abbasian et al. (2015) repor­ted that PO ₄ ^3- concentration was statistically higher in the TF beneath a beech forest located in the Caspian forests of northern Iran than in rainfall. Others have reported that TF under deciduous trees had higher PO ₄ ^3- concentrations relative to GR . For example, Rodrigo et al. (2003) showed that PO ₄ ^3- in TF under an oak stand increased after passing through the canopy mainly due to the leaching process. However, Ling-Hao & Peng (1998) showed that the canopy absorbed PO ₄ ^3- during the non-growing season in a Castanopsis eyrei stand.

The increase in Mg ²⁺ concentration in our study was in consistent with Balestrini et al. (2007) . They reported that the concentration of Mg ²⁺ increased because of interactions with the canopy in P. abies and F. sylvatica stands due to the wash off of dry deposited Mg ²⁺ from the crown surface. Dezzeo and Chàcon (2006) also showed that the concentration of Mg ²⁺ increased 3-4 fold.

Na ⁺ concentration increased after canopy leaching especially in P. abies so that the concentration of this element in SF and TF of spruce under the canopy was statistically higher compared with beech forest (Fig. 3). Lu et al. (2017) showed that the concentration of Na ⁺ in TF beneath a Pinus densata stand was more than that of GR . In addition, Parker (1983 ) noted that the annual Na ⁺ return to the forest soil predominantly via TF and SF and to a lesser extent through litterfall. In general, leaching process from canopy trees is the main reason for increasing the concentration of cations in TF compared with GR ( Balestrini et al. , 2007 ; Staelens et al. , 2007 ; Adriaenssens et al. , 2012 ). Balestrini et al. (2007) who measured the concentrations of cations input through TF in oak, European beech, and P. abies forests in Italy, reported that Ca ²⁺ , K ⁺ , Na ⁺ , Mg ²⁺ and NH ₄ ⁺ concentrations were higher in both broadleaf and conifers than GR .

Similar to Muoghalu and Oakhumen (2000) , we showed that concentrations of PO ₄ ^3- , Ca ²⁺ , K ⁺ , Mg ²⁺ , and Na ⁺ in the SF generated by both stands were more than what was found in of GR (Fig. 3). Moreover, the amounts of Ca ²⁺ and Na ⁺ elements in the spruce forest were higher than that of beech forest. This observation was consistent with the results of Houle et al. (1999) who showed that concentrations of Ca ²⁺ and Na ⁺ in coniferous forest were higher than deciduous. The PO ₄ ^3- concentration in SF of beech forest was higher than that of the spruce stand (Fig. 3). Liu et al. (2003) compared the nutrients of GR and SF in a natural mixed forest and concluded that SF had higher concentration of Na ⁺ , K ⁺ , Ca ²⁺ , and Mg ²⁺ . Dezzeo and Chacón (2006) by examining the changes in SF in a Savannah forest showed that the average concentration of nutrients in the SF were higher than those in GR .

We detected no significant difference in the concentration of NO ₃ ^- between the GR and SF and TF in both stands (Fig. 3). In addition, no significant difference was observed in NH ₄ ⁺ concentrations bet­ween SF and GR for both stands. Houle et al. (1999) state that NO ₃ ^- and NH ₄ ⁺ is absorbed by branches and trunks of deciduous and coniferous stands. Houle et al. (1999) report that a coniferous stand had higher uptake relative to a deciduous stand, in part, because of epiphytic lichens (and associated microorganisms) that grow on trunks in the coniferous stand.

Water passing through the O layer during a rain event increases the cations entering the mineral soil ( Eaton et al. , 1973 ; Bernhard-Reversat, 1975 ). For both stands, nutrients leaching from the O layer were either similar to (NO ₃ ^- , NH ₄ ⁺ ) or significantly greater than (K ⁺ , Na ⁺ , Mg ²⁺ ) the concentrations in GR . The greatest difference in nutrient fluxes between the two stands was the significantly greater fluxes of PO ₄ ^3- to the mineral soil in the beech stand. Moreover, the difference in the chemical composition of O layer in both stands and the interaction of the different elements with the O layer are considered important factors controlling nutrient fluxes from the O layer ( Hojjati et al. , 2009 ; Adriaenssens et al. , 2012 ). For example, Hojjati et al. (2009) stated that the importance of TF and litterfall fluxes in total nutrient inputs to the soil surface varies depending on the nature of the elements. Stachurski and Zimka (2002) demonstrated that nearly 80% of K ⁺ in foliage was in ionic form, higher than those for Mg ²⁺ (40%) and Ca ²⁺ (20%). In general, TF concentrations in unlike stands explains the difference in cation concentrations exciting O layer in different stands ( Ashagrie & Zech, 2010 ). In our stands, PO ₄ ^3- was significantly higher in both TF and the O layer in the beech stand.

The magnitude of the change in a particular element depends on the type of forest (coniferous or broadleaf), tree species, forest structure and other ecological and climatic factors. In addition, after the TF and SF reach the forest floor, their chemical composition changes yet again when leaching through soil organic horizons. Information on the quantity and quality of the nutrient cycle in forest ecosystems and the impact of planting exotic species on these cycles provides essential and practical knowledge for better management of these forests.

We observed a significant decrease in the pH of GR when water passes through soil litter layer in both forests. The higher LAI in the spruce stand likely contributes to the increased leaching of Ca ²⁺ , Mg ²⁺ , K ⁺ , and Na ⁺ . In contrast, the trunk and branches of the beech forest significantly increased the concentrations of PO ₄ ^3- . The differences in cation concentrations exciting the O layer appears tightly linked to changes in TF .