Introduction

In forests with symptoms of damage by air pollution, changes in nutrient foliage in different seasons could help detect early changes of nutritional status (^{López, 1993}; ^{Kurczynska, et al., 1997}; ^{Alvarado-Rosales and Hernández-Tejeda , 2002}). When some air pollutants dissolved in rainwater can generate acid attacking the leaves cuticle and accelerating nutrient washing tree top (^{López., 1993}; ^{Fenn et al., 1999}; ^{Saavedra, 2001}). Washing tree top is an important process that represents the transfer of nutrients from the canopy to the ground (^{Hamburg and Lin, 1998}). There are few studies on the chemistry of foliar runoff in forests Abies religiosa (^{Fenn et al., 1999}; ^{Saavedra et al., 2003}; ^{Fenn et al., 2006}; ^{Pérez-Suárez et al., 2006}), although some nearby forest areas to the cities, are showing effects on the foliage as a result of air pollution.

In Mexico, already have reported damage from air pollution in forests of the park Desierto of the Leones, located southeast of Mexico City (^{Alvarado, 1989}; ^{Fenn et al., 1999}; ^{Fenn et al., 2002}, ^{Fenn et al., 2006}). However, it has not found a direct relationship between air pollution damage and changes in nutritional content. In forests in decline, it is difficult to separate the effects of each factor stressed vegetation as it is a complex problem with outcomes and Synergistic starting with air pollution (^{Alvarado, 1989}; ^{López, 1993}; ^{Saavedra, 2001}).

Because fir forests located northeast of the city of Mexico are showing symptoms similar deterioration to those reported for the Desierto of the Leones (loss and discoloration of foliage and death of twigs), this work two objectives are proposed: (1) to investigate whether variations in foliage nutrient content are explained by the degree of damage in stands of fir tree top; and (2) estimate the amount of NO₃^-, NH₄⁺ and SO₄²- in the wash tree top two forests differing in proximity to Mexico City.

Materials and methods

Study zone

They were selected two forests of Oyamel state of Mexico, the first in San Miguel Tlaixpan (SMT), where they have stands with different degrees of damage tree top, and the second located in Rio Frio (RF), control site where no damage symptoms tree top were observed (Figure 1).

Figure 1 Location of the study areas. 

Features San Miguel Tlaixpan sites (SMT)

They are 45 km northeast of Mexico City and 15 km from Texcoco, State of Mexico, between 19º 18’ 25’’ y 19º 19’ 35’’ and 19 ° 19' 35 '' north latitude and between 98 ° 14 '16' 'and 98º 19 '32' 'west longitude. The altitude varies from 2 600 to 3 800 m. The weather that occurs in the area is type C (w’’²) (w) (b’)_i (^{García, 1987}), tempered humid with summer rains and annual average temperature of 13 ° C. the annual rainfall is 1 180 mm, of which 75% is concentrated between the months of June to September (^{Rey, 1975}). The soils are humic Andosol, ochric Andosol and háplico Feozem (^{FAO, 1988}), are deep soils with frank predominant texture sandy, moderately plastic, very rich in total nitrogen and organic matter, and saturation percentage lower base 50% (^{Marin et al., 2002}). In the Table 1 shows the soil determinations were performed to characterize the sites under study.

Table 1 General characteristics of the soil in San Miguel Tlaixpan and Rio Frio, State of Mexico. 

ESMT= Ejido San Miguel Tlaixpan; ERF= Ejído Río Frío. Error estándar entre paréntesis.

Results

In general, the concentration of N, P, and Mg decreased in all forest conditions during the observation period (Figure 2). Except for Ca and K, in general foliar nutrient concentration it decreased from spring to winter. However, in the healthy forest RF the concentration to Mg showed no change over time and maintained its value of 0.15%. This forest RF showed Mg content equivalent to 10% of the nitrogen content according to ^{Binkley (1993)} is a suitable value because the optimum amount for conifers Mg ranging from 4 to 9% equivalent to the content of N. The higher concentrations N were observed in the RF forest.

Figure 2 Concentration of N, P, K, Ca, Mg and Mg / Ca and N / N ratio Abies religiosa. DS = severe damage; DI = intermediate damage; S-SMT = sano San Miguel. 

Unlike other nutrients concentrations of Ca and K and Ca/N ratio increased in the four forest conditions over time. The Mg/N ratio showed a tendency to decrease; however, unlike other conditions, shows greater proportion RF forest Mg/N in the last sampling date, suggesting that there is more availability of Mg in this forest.

In the Table 2 shows the results of statistical analyzes. The effect of forest condition by time interaction was not significant for N, K, Ca, and Ca/N ratio. However, this same trend is observed for P and in this case if the differences are significant. This, by the similar trend (parallel) which showed these variables over time is explained. Time interaction was significant condition for P, Mg, and the Mg/N, indicating differences in concentration between the different forest conditions over time. Unlike other conditions, RF shows higher values of the ratio Mg / N in the last sampling date.

Table 2 Results of statistical analysis of foliar concentrations with respect to their status, time and interaction. 

With respect to NO₃^-N deposit no significant differences between the SMT and RF sites were found. However, NH₄⁺ -N differences and SO₄² -S were, showing greater amounts in SMT (Figure 3).

Figure 3 Flow N-NO₃^-, N-NH₄⁺ and S-SO₄^2- tree top washing two fir forests of Rio Frio and San Miguel Tlaixpan, State of Mexico. Annual data are estimated from sampling period July to January and performing a proportional adjustment for the historical annual rainfall. 

Discussion

The concentrations of N, P and Mg found in the spring season, they are higher than those reported by ^{Alvarado (1989)}. However, during the summer and fall, they are comparable with those found by this author. In Zoquiapan (ZQ) and Desierto of the Leones (DL) fluctuations reported by ^{Alvarado (1989)} were 0.91-1.29%, 0.08 to 0.15%, and 0.08 to 0.09% for N, P and Mg, respectively. Fluctuations in this work for the fall term were 0.09-1.2%, 0.07 to 0.8% and from 0.1 to 0.17% for N, P, and Mg respectively.

The results show that the N, P, and Mg are elements that reduce its concentration in plant tissue as the growing season progresses and remains relatively stable in the summer and fall. Higher concentrations of N and Mg were observed in RF in summer and winter. The results of nutritional analysis conditions not statistically separated damage tree top used in this study. The K concentrations found in the spring period in this study are lower than those reported by ^{Alvarado (1989)}, ^{Lopez (1993)} and ^{Saavedra (2001)}. However, those found for periods of summer and autumn, are greater than those indicating ^{Saavedra (2001)}. For DL region ZQ and ^{López (1993)} reports an average K value of 0.79%; ^{Saavedra (2001)} reports an average value of 0.446% and ^{Alvarado (1989)} of 0.55-1.58%. In this study fluctuations for the period from spring to autumn were 0.20 to 0.90%.

The results indicate that the concentration of nutrients depends on the time of collection of the foliage, so that comparisons of nutrient concentration between similar studies are difficult. The K results did not vary with damage tree top and also failed to statistically separate forest conditions considered. ^{Saavedra (2001)}, also found no significant differences in the concentration of K between healthy and diseased trees in the region of Desierto of the Leones (DL). According to ^{Lopez et al. (1998)}, the K is one of the weakest elements in natural forest of Abies religiosa; however, for the three forest conditions considered in this study show no trend regarding this item.

With respect to Ca concentrations spring period were lower than those reported by ^{Alvarado (1989)} for the DL and ZQ region where the variation was 0.88-1.94%. Fluctuations in this work for the fall term were 0.20-1.9%. Ca interval is broader than reported by other authors, possibly by considering the latent and tree growth periods. The trend in the increase of Ca over time is expected since the Ca is an important structural component of the cell wall and middle lamella (^{Binkley, 1993}).

In studies with Pinus sylvestris have found higher foliar concentrations of K, Ca, and Mg (^{Kurczynska et al., 1997}) in forests growing on contaminated atmospheres, compared with witnesses forests. ^{Whytemare et al. (1997)} studied Picea glauca forests found in areas under air pollution was higher; however, foliar Ca and Mg concentrations were lower compared to the control forests. According to ^{Marschner (1990)}, when K⁺ is poor tree growth slows by which this nutrient is retranslocated of mature leaves and stem. However, if the deficiency is severe and persistent these organs may become chlorotic or necrotic. Around this, it is possible that low concentrations of K⁺ in the foliage of two years reflect part or all of any kind of stress. Values of the Mg/N and Ca/N obtained are higher than those reported by ^{Whytemare et al. (1997)}, who found that contaminated areas near the Mg/N was 0.02 to 0.04% for Picea glauca; while the Ca/N ratio in low pollution areas was 0.01 to 0.02%. In this study the minimum values of the Mg/N y Ca/N were 0.2% in both cases. This result suggests that no significant nutritional imbalances are reflected among the different degrees of damage tree top.

The N-NO₃^- deposit found in this work ranged from 2.84 to 4.40 kg ha^-1 yr^-1 for RF and SMT, respectively. Likewise, variations of N-NH₄⁺ were 4.34 to 8.56 kg ha^-1. These results are interpreted as moderate deposits, considering the work of ^{Fenn and Poth (2004)}, who found that in a forest of Pinus ponderosa low pollution, the amounts deposited were 9.1 and 7.9 kg ha^-1 for N-NO₃^- and N-NH₄⁺. However, for the case of N-NH₄⁺ differences it was statistically significant between sites, indicating greater washing tree top N as ammonium in SMT. The total deposit of N in SMT, which is the site problem was 12.96 kg ha^-1 yr^-1, which is also described as a moderate load when compared to the work of ^{Fenn et al. (1999)} and ^{Fenn et al. (2002)}, who reported a total amount of N laundering tree top of 18.5 kg ha^-1 for the DL forest is adjacent to Mexico City.

However, the total amount of N found in SMT, is greater with the amount reported for a forest of Pinus Hartwegii, used as a control where the total N deposition washing tree top was 5.5 kg ha^-1. The total amount of N for RF was 7.18 kg ha^-1 yr^-1, which is higher than reported for the control of forest ^{Fenn et al. (1999)} and located in the experimental field Zoquiapan, Mexico. Importantly, in the case of measuring ZQ tree top washing it was by conventional methods, while this study were based collectors with ion exchange resin. The annual load of N found in SMT is comparable with showing some conifers forests adjacent to the city of Mexico (^{Perez et al., 2008}; ^{Silva et al., 2015}.). It is also important to mention that this work is not considered stemflow, so the deposit of ions studied could be higher.

^{Butler and Likens (1995)} found that the amounts of wash tree top total N, including runoff per trunk was 12.4 kg ha^-1 yr^-1 in a deciduous forest composed of Quercus rubra, Quercus alba, Fagus grandifolia and Acer saccharum. The authors considered organic inorganic total nitrogen (NO₂⁺ NH₄⁺) and. Therefore, the results of this study are relatively moderate since the total sum of NO₃^-+ NH₄⁺ was 12.96 kg ha^-1 in SMT, which is the site with the highest deposition. The results found in this study are comparable to those in the study of ^{Arthur and Fahey (1993)} who when considering only station liquid rain measured 6.0 kg N ha^-1 yr^-1 in a forest of Picea engelmannii and Abies laciocarpa in the state Colorado in the United States. It is inferred that in considering the season with snow, results for the study of ^{Arthur and Fahey (1993)} are higher and therefore the degree reaffirms low relative income of N in the forests studied.

Compared with other studies the results of this work are low; for example ^{Lin et al. (2000)} they found in a subtropical forest streams washing tree top equivalent to 16, 17 and 80 kg ha^-1 yr^-1 for NH₄⁺, NO₃^-, SO₄²- and respectively. For European forests (Netherlands) deposition in forests with moderate pollution is 17, 34 and 59 kg ha^-1 yr^-1 for NH₄⁺, NO₃^-, SO₄²- and respectively. There deciduous forests where the deposition is very moderate; for example, the growing season for some deciduous forests in Douglas County, Kansas, USA, reported deposition is lower than that found in this work with values of 0.27, 1.0 and 2.9 for NH₄⁺, NO₃^-, SO₄²- and respectively. Although these values correspond only to the growing season, most rainfall occurs during this period, so it would be expected that the annual values do not exceed those reported in this study.

The deposition of inorganic N is influenced by the proximity to anthropogenic sources, ^{Fenn et al. (2004)} reported similar depositions of 0.7 kg ha^-1 yr^-1 of N as nitrate and ammonium for forests of northern California, while forests of San Bernardino near the city of Los Angeles has depositions of 53 and 45 kg N in the form of ammonium and nitrate respectively.

The deposit SO₄^2- is also relatively high considering the measurements reported by ^{Fenn et al. (2002)} who found a deposit of 8.8 kg ha^-1 in a forest of Pinus hartwegii used as control. The same authors report SO₄^2- deposits in problem sites as the site DL up to 20.4 kg ha^-1. Butler and Likens (1994) found that the amounts of wash tree top, including runoff per trunk was 14.9 kg^-1 S has a deciduous forest of Ithaca, NY. However, it should be noted that the main difference in precipitation due to sulphates, where significance was greater than 0.01. Although the amounts of ions are low compared to other studies, a clear difference between the two forest sites found.

Although the chemical composition of tree top washing is affected by plant species and proximity to anthropogenic sources, some authors have suggested that the critical load of SO₄^2- is 25 mmol m^2-1 year^-1. The load encountered at the site of SMT, the highest deposition, corresponds to 15.5 mmol m^2-1 year^-1. Making estimates for locations such as DL (^{Fenn et al., 1999}), the burden of sulphates is 21 mmol m^2-1 year^-1, a value close to the critical load. These data indicate that the SMT site is not as pristine as it was thought.

The proportion SO₄^2-/NO₃^- is an important chemical characteristic of the washing solution of the tree top , ^{Neary and Gyzin (1994)} measured proportions of 2.5 kg ha^-1 in a forest consisting of Pinus strobus, Tsuga canadensis, Acer rubrum, Quercus rubra and Betula papyrifera, Ontario Canada. In this study a ratio of 2.3 was found and 4.1 kg ha^-1 for locations SMT and RF, respectively. However, this relationship is very variable for each forest type and location, for example in the work of ^{Whelan et al. (1998)} this index reaches 30. The results of this study indicate dominance SO₄^2- is an ion that can lead to acid rain and soil solution.

However, this effect was seen in the results of soil analysis, probably because the soil buffering capacity is large enough to resist pH changes (^{Fenn et al., 2006b}). Another reason is that the acidity of rain at critical sites of the Valley of Mexico have not reached acidic values (pH less than 5.0) (^{Saavedra et al., 2003}). A potential indicator of acidity is the mass corresponding to NH₄⁺, NO₃^-, SO₄^2- and (^{Erisman et al., 2001}). Using the potential acidity, met values of 536 and 266 mol ha^-1 yr^-1, which are comparatively low values with European forests (Netherlands) where the potential acidity ranges from 2800 up to 5 000 mol ha^-1 yr^-1 (^{Erisman et al., 2001}).

Conclusions

The separation of the three conditions of tree top damage according to their nutritional content was unclear leaf except for the witness RF forest. The witness RF site, used as a control in this study showed better levels of N and Mg foliar mainly in the summer and winter period. It was found that washing the tree top of SO₄^2- ions and NH₄⁺ was higher in San Miguel Tlaixpan compared with Rio Frio. However, the two sites critical samples indications N deposit levels, which can be related to the quality of tree top in these places.