- Email: [email protected]
Artificial recharge of groundwater through sprinkling infiltration: Impacts on forest soil and the nutrient status and growth of Scots pine
Science of the Total Environment 407 (2009) 3365–3371
Contents lists available at ScienceDirect
Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v
Artificial recharge of groundwater through sprinkling infiltration: Impacts on forest soil and the nutrient status and growth of Scots pine Pekka Nöjd a,⁎, Antti-Jussi Lindroos a, Aino Smolander a, John Derome b, Ilari Lumme a, Heljä-Sisko Helmisaari a a b
Finnish Forest Research Institute, Vantaa Research Unit, P.O. Box 18, FI-01301 Vantaa, Finland Finnish Forest Research Institute, Rovaniemi Research Unit, P.O. Box 16, FI-96301 Rovaniemi, Finland
a r t i c l e
i n f o
Article history: Received 31 October 2008 Received in revised form 27 January 2009 Accepted 28 January 2009 Available online 9 March 2009 Keywords: Artificial recharge Forest soil Groundwater N transformations Needle nutrients Radial growth Pinus sylvestris (L.)
a b s t r a c t We studied the chemical changes in forest soil and the effects on Scots pine trees caused by continuous sprinkling infiltration over a period of two years, followed by a recovery period of two years. Infiltration increased the water input onto the forest soil by a factor of approximately 1000. After one year of infiltration, the pH of the organic layer had risen from about 4.0 to 6.7. The NH4-N concentration in the organic layer increased, most probably due to the NH4 ions in the infiltration water, as the net N mineralization rate did not increase. Sprinkling infiltration initiated nitrification in the mineral soil. Macronutrient concentrations generally increased in the organic layer and mineral soil. An exception, however, was the concentration of extractable phosphorus, which decreased strongly during the infiltration period and did not show a recovery within two years. The NO3-N and K concentrations had reverted back to their initial level during the two-year recovery period, while the concentrations of Ca, Mg and NH4-N were still elevated. Nutrient concentrations in the pine needles increased on the infiltrated plots. However, the needle P concentration increased, despite the decrease in plant-available P in the soil. Despite the increase in the nutrient status, there were some visible signs of chlorosis in the current-year needles after two years of infiltration. The radial growth of the pines more than doubled on the infiltrated plots, which suggests that the very large increase in the water input onto the forest floor had no adverse effect on the functioning of the trees. However, a monitoring period of four years is not sufficient for detecting potential long term detrimental effects on forest trees. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Shortage of good quality drinking water is an increasing problem in many parts of the world and techniques for producing safe, pure water are therefore highly valuable. In many EU countries artificial recharge of groundwater (AR) is an important means for reaching the goals of the Water Framework Directive. In Finland, 60% of the household water distributed by the waterworks is either natural or artificially recharged groundwater produced by basin and sprinkling infiltration. The proportion of AR groundwater out of total household water production, which is currently ca. 15%, is expected to increase during the next decade. Tampere and Turku, the third and fifth largest cities in Finland, are in the process of adopting the technique. AR groundwater is also widely used in, for example, Sweden. During the production of AR, the concentrations of particulate and dissolved organic carbon in the surface water are reduced by filtration through the overlying soil (e.g. Frycklund, 1995; Lindroos et al., 2002). Chemical and biological processes in the soil also contribute to the reduction in organic carbon. ⁎ Corresponding author. Tel.: +358 10 211 2331; fax: +358 10 211 2203. E-mail address: pekka.nojd@metla.fi (P. Nöjd). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.01.062
Most of the Finnish AR groundwater plants are situated on forested eskers formed at the end of the last ice age. Basin infiltration has been used for decades in several cities for producing AR. Recently, however, many of the eskers have been designated as protected areas or their use has been restricted. Therefore, as the excavation of basins destroys the natural vegetation and topsoil, the establishment of new basins in esker areas is faced with conflicting interests. Since the mid-1990's, Finnish cities have also used a new method, sprinkling infiltration, for AR purposes. In the method, lake or river water is sprinkled directly onto the forest floor from a network of pipes without removing the ground vegetation or the top soil (Helmisaari et al., 1998) (Fig. 1). As no basins are needed, sprinkling infiltration reduces the direct physical disturbance of the esker ecosystems. Sprinkling infiltration has been in use at Jyväskylä, a city with a population of 85 000, since the late 1990s. The artificial recharge of groundwater for supplying the water requirements of large cities involves huge quantities of infiltration water. As the sprinkling infiltration areas are relatively small, the input of water per unit ground surface may be thousands of times higher than the annual precipitation. In addition, the chemical composition of the water used for infiltration differs from that of precipitation. The water source used for producing artificial groundwater for Jyväskylä, Lake Kuusvesi, is a
3366
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
Fig. 1. The principle of producing artificially recharged groundwater (AR) by sprinkling infiltration.
fairly large (22 km2) unpolluted, oligotrophic lake. Its water has a higher content of total organic carbon (TOC), 6.7 mg/l, and a higher mean pH (pH 7.0) than precipitation. It is therefore important to monitor and study the changes in forested esker ecosystems caused by infiltration, and to investigate the possibilities of minimizing the potential adverse effects of such disturbances. The extremely large amounts of water added continuously during sprinkling infiltration have been shown to alter the fluxes of water and nutrients in the soil, resulting in an increase in the pH and base cation concentrations of the uppermost soil layers (Lindroos et al., 2001), and changes in nitrogen cycling (Paavolainen et al., 2000b) and the species abundance of the understorey vegetation (Helmisaari et al., 1993). The results from a study on a fertile forest site in southern Finland (Helmisaari et al., 1998) resulted in recommendations that each site should be infiltrated for a maximum of two growing seasons, followed by a break without irrigation lasting for at least one growing season. In 1998, we initiated a research project at Vuonteenharju, Central Finland, where the water works of the city of Jyväskylä had installed a sprinkling infiltration plant on a nutrient-poor forest site. We monitored the stability of the ecosystem on the forested esker during and after a two-year infiltration period. In this article we report the effects of sprinkling infiltration on soil chemistry, soil nitrogen transformations, and the nutrient status and radial growth of trees. The recovery of the ecosystem was also evaluated. 2. Materials and methods 2.1. The study site The studied infiltration area at Vuonteenharju, Central Finland, is located on an esker 18 km north-east of Jyväskylä. Since 2000, 70% of the household water for the city's population of 85 000 has been produced by artificial recharging through sprinkling infiltration at Vuonteenharju. The infiltration area is located in a Scots pine (Pinus sylvestris) stand on a sub-xeric site representing the Vaccinium vitis-idaea forest site type (Cajander, 1949). The soil consists of relatively coarse, stratified sand deposits. The soil type is humic podzol. Four experimental plots were established in 1998, each 30 × 30 m in size. Two of the plots were infiltrated during 20.9.1999−19.12.2001, and two remained as untreated controls. In addition, another area within the same forest stand, which was infiltrated during 16.11.2002−2.5.2005, was used for studying the effect of infiltration on the diameter growth of Scots pine. The forest had been regenerated during the 1890s. The stand was fully stocked (770 stems/ha, cubic volume 294 m3/ha, dominant height 22.3 m), and had not been treated to silvicultural thinnings for
several decades. The dominant height of 22.3 m at the age of more than 100 years reflects medium-to-low productivity. 2.2. The infiltration treatment During sprinkling infiltration, the annual amount of infiltrated water was c. 600 m3 m− 2, which is about 1000 times the water input from precipitation in this area (Helmisaari et al., 2003). Although the median nitrogen concentrations (NH4-N b0.002 mg/l, NO3-N 0.07 mg/l and Ntot 0.35 mg/l) of the infiltrated lake water were relatively low, the annual amounts of nitrogen added to the forest floor were considerable due to the very large amounts of infiltrated water. A rough estimate of the input of Ntot into the forest floor exceeds 2000 kg/ha/yr and the input of NO3-N was N400 kg/ha/yr, while the magnitude for NH4-N was estimated to be several kilograms/ha/yr. 2.3. Soil samples In order to study the effects of infiltration on soil nitrogen transformations (net N mineralization and nitrification), as well as on factors related to these processes, 18 soil cores (diameter 25 mm) were taken from the organic layer before the start of the treatment (September 4, 1998) and one month (October 25, 1999), one year (September 25, 2000) and two years (September 18, 2001) after the infiltration started. The soil cores were taken systematically at a distance of one meter from the sprinkling infiltration pipes, and bulked to give one composite sample per plot. This means that all the sampling points were subjected to the effects of irrigation. In 2001 samples were also taken from the uppermost 5 cm of the mineral soil layer. The control plots were sampled in a similar fashion. The fresh soil samples were sieved (mesh size 2.8 mm). Soil pH was measured in a water suspension, and the concentrations of NH4-N and NO3-N after KCl extraction as described in Paavolainen et al. (2000b). The rates of net N mineralization and net nitrification were measured in incubation experiments at constant temperature and moisture (Paavolainen et al., 2000a). In order to study changes in the nutrient status (extractable NH4, NO3 and P and exchangeable Ca, Mg and K concentrations) of the topsoil, as well as possible recovery in the nutrient status after infiltration had ceased, soil samples were taken in 2000, 2001, 2002 and 2003 from both the infiltrated and control plots. The organic layer and mineral soil layers (0–5, 5–10, 10–20 and 20–40 cm depth) were sampled at 20 systematically located points on each plot. All the points sampled on the plots were subjected to the effects of irrigation. The samples were combined to give two samples per layer for each plot, and dried at 60 °C. The organic layer samples were milled to pass through a 1 mm sieve, and the mineral soil samples passed through a 2 mm sieve. The extractable NH4 and NO3 concentrations were
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3367
Table 1 pH, concentrations of NH4-N and NO3-N and rates of net N mineralization and net nitrification in the humus layer of the infiltrated and control plots.
pH
NH4-N, mg/kg o.m.
NO3-N, mg/kg o.m.
Net N mineralization (NH4 + NO3)-N, mg/kg o.m./6 wk
Net nitrification NO3-N, mg/kg o.m./6 wk
1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001
Control
Infiltration
3.9 4.0 3.7 3.7 113.8 2.3 17.5 10.9 0.0 0.0 0.6 0.0 − 101.3 5.2 21.1 − 5.3 0.0 0.2 − 0.1 0.0
3.8 5.6 6.7 6.7 112.4 12.1 170.6 117.3 0.0 0.0 11.9 5.5 − 103.4 0.5 − 63.2 − 96.6 0.0 0.3 2.6 − 2.6
Mean for the two replicate study plots. o.m. = organic matter.
Fig. 2. The difference between the infiltrated and control plots in NO3-N concentration (mg/kg of dry matter) one year (2002) and two years (2003) after the termination of infiltration in (a) the organic and (b) the uppermost mineral soil layers.
each sample tree in September 2006. The ring widths were measured to an accuracy of 0.01 mm, and cross-dated using COFECHA software (Holmes, 1983). 3. Results 3.1. Effects of infiltration on soil nitrogen transformations
determined after extraction with 1 M K2SO4 (2 g organic layer or mineral soil/50 ml extractant) by flow injection analysis (FIA). Exchangeable Ca, Mg, K and extractable P concentrations were determined after extraction with 0.1 M BaCl2 (3.75 g organic layer or 15 g mineral soil/150 ml extractant) by inductively coupled atomic emission spectrometry (ICP/AES). 2.4. Needle samples Needle samples were taken from the pines growing close to the infiltration pipes before the treatment (1.10.1998) and after two years of infiltration (18.9.2001), and again from the same trees two years after termination of the treatment (22.9.2003). The radial growth of five of the eight trees sampled per plot was also measured. Two branches from the upper third of the canopy were sampled on each tree. The samples were stored frozen for up to 20 days, the needle age classes separated, and the needles then dried (+50 °C, 72 h) and milled. Total nitrogen was determined on a LECO CHN analyzer, and the P, K, Ca, Mg, Mn, Mo, Zn, Fe, Cu, Al, Ni and B concentrations by ICP/AES after dry digestion (550 °C for 2 h) and extraction with HCl.
The pH of the organic layer was initially slightly below 4 (Table 1), but after one year's infiltration it had risen to 6.7. After both one and two years of continuous sprinkling infiltration, the NH4-N concentration in the organic layer increased (Table 1). The NO3-N concentrations were initially negligible, as well as after one month of sprinkling infiltration. Towards the end of the infiltration period there was a clear increase in the NO3-N concentrations. Net N mineralization was negative before the onset of sprinkling infiltration, indicating either a low rate of gross N mineralization or very high immobilization of nitrogen. During the course of sprinkling infiltration there was an apparent decrease in net N mineralization. Net nitrification was negligible before the onset of sprinkling infiltration, and there was no consistent effect during sprinkling infiltration. In samples taken from the uppermost mineral soil (in 2001) after two years of continuous sprinkling infiltration, net nitrification was negligible on the control plots, but intensive on the sprinkling infiltration plots (Table 2).
Table 3 The Ca, Mg, K, P, NH4-N and NO3-N concentrations (mg/kg d.m.) in the organic layer and mineral soil at depths of 0–5, 5–10, 10–20 and 20–40 cm.
2.5. Growth measurements A total of 22 pines were cored in autumn of 2006 in order to study the effect of infiltration on radial growth. Ten trees were cored on the infiltrated plots (infiltration between 20.9.1999−19.12.2001). Another seven trees were selected from the area infiltrated during 16.11.2002−2.5.2005. Five pines outside the sprinkling infiltration areas, but within 50 m of these two areas, were cored as control trees. An increment core from the bark surface to the pith was taken from
Soil layer
Treatment
Ca
Mg
K
P
NH4-N
NO3-N
Organic layer
Control 2002 — Infiltr. 2003 — Infiltr.
2516⁎ 8033⁎ 7605⁎
308⁎ 1030⁎ 2198⁎
694⁎ 625⁎ 506⁎
143⁎ 53⁎ 55⁎
34.10 51.50 80.70
0.00 0.66 0.31
Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr.
67.5 311.0 402.0 22.4 126.0 173.0 15.0 59.6 62.3 9.6 31.1 42.4
10.70 45.30 53.70 4.20 18.00 22.80 3.09 9.22 10.17 2.12 5.23 7.50
24.90 21.30 24.70 15.20 12.00 16.90 12.50 6.81 7.95 6.10 6.83 9.85
12.20 0.85 0.73 11.70 0.40 1.90 10.50 0.51 1.03 8.20 0.51 2.40
1.84 2.53 2.22 1.84 2.60 2.24 1.84 2.32 1.38 1.84 1.86 1.04
0.00 0.14 0.00 0.00 0.20 0.00 0.00 0.32 0.00 0.00 0.39 0.00
Mineral soil 0–5 cm
5–10 cm
Table 2 pH, concentration of NO3-N and rate of net nitrification in the uppermost mineral soil in 2001.
pH NO3-N, mg/kg d.m. Net nitrification, mg/kg d.m./6 wk Mean for the two replicate study plots. d.m. = dry matter.
Control
Infiltration
4.5 0.0 0.0
6.3 0.5 11.4
10–20 cm
20–40 cm
Control = mean for the control plots without infiltration, 2002 = mean for the infiltrated plots one year after the cessation of the infiltration, 2003 = mean for the infiltrated plots two years after the termination of the infiltration. Values marked with ⁎ are from Derome et al. (2006).
3368
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3.2. Ammonium, nitrate, phosphorus and base cation concentrations in the soil after the termination of infiltration
Fig. 3. The difference between the infiltrated and control plots in Ca concentration (mg/kg of dry matter) one year (2002) and two years (2003) after the termination of infiltration in (a) the organic and (b) the uppermost mineral soil layers.
The NO3-N (Fig. 2) and NH4-N concentrations in both the organic and mineral soil layers on the infiltrated plots were still higher than on the controls one year (2002) after the treatment had ceased, and the NH4-N concentration in the organic layer was even higher after two years (Table 3). The Ca (Fig. 3) and Mg concentrations in both the organic and mineral soil layers on the infiltrated plots were still higher than on the control two years after the termination of infiltration (in 2003) (Table 3). Neither infiltration nor the termination of infiltration had any effect on the K concentrations. Infiltration caused a strong decrease in the extractable P concentrations, and no clear recovery was observed within two years after the termination of infiltration (Table 3).
Fig. 4. Nutrient concentrations in current year (C) and one-year-old (C + 1) needles of Scots pine sample trees on the control and infiltrated plots before infiltration in 1998, after two years of infiltration in 2001, and two years after the termination of infiltration in 2003 (mean and s.d. of 16 trees per treatment).
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3.3. Tree nutrient status Before the start of infiltration there were only slight differences in needle nutrient concentrations between the control and infiltrated plots, but two years of sprinkling infiltration caused significant changes in the needle nutrient status. Especially the concentrations of N, but also of P, K and B, had increased in the needles on the infiltrated plots by 2001 (Fig. 4). Two years (in 2003) after the end of infiltration, the N, P, K and B concentrations in current-year needles (formed in 2003) had returned to the level prior to infiltration. In contrast, the Ca and Mg concentrations reached their highest levels two years after the termination of infiltration. The nutrient status of the one-year-old needles (formed in 2002) closely followed the patterns in the currentyear needles (Fig. 4). Micronutrient concentrations in the needles (results not shown) did not indicate any signs of nutrient deficiencies during and after infiltration. 3.4. Radial growth Prior to the onset of infiltration, the growth of the sample trees from the two infiltrated areas and the control followed a very similar pattern (Fig. 5a). In 1981, an outbreak of Neodiprion sertifer, an insect that feeds on pine needles, caused a rapid reduction of growth, followed by a gradual recovery during the following 7−8 years. After the onset of infiltration, the mean ring width increased from 0.57 mm in 1999 to 1.47 mm in 2002 (Fig. 5b) on the plots infiltrated
Fig. 5. The mean annual radial growth of Scots pines in the area infiltrated during 20.9.1999–19.12.2001 (red line), in the area infiltrated between 16.11.2002–2.5.2005 (blue line) and in the control (yellow line). The periods shown are 1910–2006 in 5a and 1990–2006 in 5b (infiltration periods shown with thicker lines).
3369
during 20.9.1999−19.12.2001. In 2006 the mean ring width decreased to 0.65 mm, close to the level prior to infiltration. On the area infiltrated between 16.11.2002 and 2.5.2005, there was a strong increase in ring-width in 2004. In 2006 the growth was still higher than that prior to onset of the treatment. The ring widths of the control trees varied only slightly during the period 1999−2006. The year 2000 was an exception, presumably due to favourable weather conditions during the growing season in Central Finland. 4. Discussion We studied the chemical changes on forest soils and the effects on forest trees caused by continuous sprinkling infiltration for a period of more than two years. The experiment consisted of two infiltrated 30 × 30 m plots and two controls of similar size. The experimental design is unsuitable for statistical comparison between treatments due to the lack of sufficient number of replications. A relatively large plot size was considered necessary for preventing surface flow from infiltrated plots to control plots. Given the large plot size, the size of the forest stand and local topography restricted the number of sufficiently homogenous study plots. In addition, as the treatment involved continuous infiltration over a period of more than two years, there would have been technical problems in launching and managing a more complex study design — a side effect of studying the effects of an operating recharge plant. In boreal forest ecosystems, the availability of nitrogen is a key factor limiting tree growth (e.g. Kukkola and Saramäki, 1983; Smolander et al., 2000). This is especially the case on nutrient-poor sites, such as the forested esker at Vuonteenharju used for producing household water by the artificial recharge of groundwater. However, while additional nitrogen is generally beneficial for trees, the excessive total amount of N in the infiltration water could alter the nitrogen cycle (Paavolainen et al., 2000b), resulting in the formation of NO3-N, and thus have highly detrimental effects on the quality of the groundwater. The NO3-N concentrations in percolation water on a fertile site in southern Finland after the termination of infiltration were N10 mg/l (Paavolainen et al., 2000b). We measured both the concentration of mineral N and the rate of net N mineralization, which depicts the amount of easily mineralizable N. The NH4-N concentration in the organic layer of the infiltrated plots increased strongly, most probably due to the NH4 in the sprinkling infiltration water because net N mineralization did not increase (Table 1). Sprinkling infiltration initiated net nitrification in the uppermost mineral soil layers (Table 2), but there was no consistent effect in the organic layer. Both NH4 availability and soil pH regulate nitrification (Smolander et al., 2000). Sprinkling infiltration provides both extra NH4 and a pH that is more optimal for nitrification. In undisturbed, acidic boreal forest soils net nitrification is usually negligible (Smolander et al., 2000). Sprinkling infiltration may increase net nitrification, but the response is apparently dependent on the site characteristics. In a study carried out on a highly fertile forest site in southern Finland, sprinkling infiltration initiated intensive net nitrification in the organic layer soon after the onset of infiltration and also continued at a high rate (Paavolainen et al., 2000b). The main reason for this was the increase in pH (Paavolainen et al., 2000a). At the less fertile site in our study, the low net nitrification in the organic layer was either due to low gross nitrification or intensive immobilization of NO3-N. A lack of substrate could not account for this phenomenon because there was an excess of NH4-N. Interestingly, the conditions for nitrification were apparently more optimal in the mineral soil. One possible reason for the lower net nitrification in the organic layer than in the mineral soil could be higher immobilization of NO3-N in the organic layer. It is also possible that the organic layer contained compounds that inhibited
3370
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
gross nitrification. For instance, volatile monoterpenes have been shown to decrease both the net N mineralization and net nitrification in the organic layer of boreal forests (Paavolainen et al., 1998; Smolander et al., 2006). The elevated NO3-N concentrations in the organic and mineral soil layers lasted for only a relatively short period of time. After a two-year recovery period the differences between the treatment and control were small. This is probably due to the fact that the trees and vegetation effectively utilized the nitrogen input: radial growth of Scots pine more than doubled on the infiltrated areas. However, the NH4-N concentrations were still elevated in the organic layer and the uppermost mineral soil layers after the two-year recovery period. Infiltration caused a clear increase in the exchangeable Ca and Mg concentrations in the organic and mineral soil layers. A similar effect has been reported previously for more fertile sites (Lindroos et al., 2001). The total input of Ca and Mg in infiltration water was extremely large and these cations are readily retained on cation exchange sites on soil particles (primarily organic) in both the organic and mineral soil layers. Neither the Ca nor Mg concentrations showed a decline during the two-year recovery period. The K concentrations in the soil showed only very small changes both during and after infiltration, probably because K ions cannot compete with Ca and Mg ions for cation exchange sites (Bohn et al., 1985; Lindroos et al., 2001; Derome et al., 2006). Infiltration caused a strong decrease in the extractable P concentrations in both the organic and mineral soil layers, and there was no recovery during the two-year period following the termination of infiltration. The relatively high pH most probably caused the immobilization of P, as reflected in the low extractable P concentrations. If the soil pH remains at a relatively high level for many years after the termination of infiltration, as has been reported for a more fertile site in southern Finland (Helmisaari et al., 2003), where no clear recovery was observed within 5 years, then the extractable P concentrations may remain at a low level for many years in our study area owing to the strong link between pH and extractable P concentrations. In boreal forests P has been shown to limit tree growth on the most fertile sites, where nitrogen deficiency is a less severe constraint (Kukkola and Saramäki, 1983). The elevated needle nutrient concentrations after two years of infiltration reflected the increased availability of nitrogen and base cations in the soil (see also Derome et al., 2006). The strong decrease in the extractable P concentrations in the soil on the infiltrated plots was not reflected in the needle P concentrations, which in fact continued to increase after the termination of infiltration. Within two years, the needle nutrient status had almost returned to the level before infiltration. After two years of infiltration, i.e. in autumn 2001, there were, however, some visible symptoms in the needles indicating possible imbalances between certain nutrients. The current-year needles (C) had chlorotic symptoms, and the needles formed during the previous summer (C + 1) had an unusually light green colour. However, this phenomenon was temporary: current-year needles formed and sampled in 2003, i.e. two years after the termination of infiltration, had a healthy green colour. The chlorosis symptoms may be related to an imbalance between nitrogen and magnesium. Magnesium, which is an important constituent of chlorophyll, was the only nutrient that showed a delayed response to infiltration. In 2001, after two years of infiltration, the Mg/N ratios were 0.046 and 0.044 in the C and C + 1 needles, respectively. In 1999 and 2003 they varied between 0.066 and 0.073 in the same trees. Braekke (1994) considered a Mg/N ratio of 0.04 in current-year needles to be a critical value. We conclude that even though the changes in needle nutrient concentration were mostly positive, some detrimental changes (e.g. needle discoloration) occurred. They, however, disappeared rapidly after the infiltration ceased. Sprinkling infiltration doubled the radial growth of the pines at Vuonteenharju. Based on our results, it is not possible to distinguish
between the growth-enhancing effects of an increased water supply and the increase in the nutrient supply to the trees. Some indications are given by Linder (1987), who studied the effects of irrigation and fertilization on the growth of Scots pine in Central Sweden (60° 49' N) during 1974–1984. After 11 years of treatment, the plots which had been treated with both irrigation and fertilization had the highest basal areas. Compared to the combination of irrigation and fertilization, the basal area on the fertilized plots was 84%, on the irrigated plots 49% and on the control plots 38%. Thus, while irrigation produced a slight response, the growth-enhancing effect of nitrogen fertilization was clearly more pronounced. Bergh et al. (1999) obtained volume increments up to four times higher than on the control plots through optimized fertilization and irrigation in two Norway spruce stands in Sweden. At their northern study site (64° 07 'N, 310−320 m a.s.l.), irrigation alone had no effect on volume growth. At the southern experimental site (57° 08 'N, 225−250 m a.s.l.), irrigation also had no significant effect on growth during the first 4-year period. Thereafter, however, the combination of fertilization and irrigation produced a 50% higher volume growth than the fertilized plots without irrigation. In both of the studies, fertilization thus induced a considerably larger growth responses than irrigation. At the northern site of Bergh et al. (1999), representing a slightly higher latitude than that of our study site (62° 20 'N), no irrigation effect was observed. Thus, the large increase in radial growth observed at Vuonteenharju can most probably be attributed to the improved nitrogen supply, which was also reflected in the elevated needle nitrogen concentrations. However, the amounts of infiltration water applied at Vuonteenharju were extremely high compared to those added by Linder (1987) and Bergh et al. (1999). Increased tree growth suggests that the extremely large input of water through the topsoil did not damage the functioning of the trees. This is further supported by the improved supply of the most important macronutrients for trees, as shown by the needle chemistry results. However, there were also observations indicating that caution is required. The discoloration of needles, faint indications of nutrient imbalances in needles, decreased extractable P concentration in the soil and severely altered soil conditions are factors suggesting that tree vitality could be endangered on the long term; especially, if the treatment is repeated on the same site without sufficient recovery periods. 5. Conclusions Virtually continuous infiltration over a period of more than two years caused clear changes in the chemical properties of the organic and uppermost mineral soil layers. The changes were mainly beneficial from the point of view of the nutrient supply of the trees. One notable exception, however, was the extractable phosphorus concentration, which decreased during infiltration and did not recover during the first two years after the termination of infiltration. Needle chemistry analyses showed that the concentrations of important tree nutrients had increased on the infiltrated plots, including the phosphorus. The needle analyses showed no clear signs of nutrient imbalances, although the Mg/N ratio of the needles was close to the critical level. The strong enhancement of radial growth also suggests that continuous sprinkling infiltration over a period of two years did not have any detrimental effects on the vitality of the pines. However, some discoloration of current-year needles, which is a symptom of nutrient imbalances, appeared immediately after the two-year infiltration period. Thus, while we found no direct evidence of detrimental effects on pine following the two-year infiltration period, there were signs of potential risks to the forest ecosystem. These features should be taken into account when planning sprinkling infiltration timetables for the production of artificially recharged groundwater. It should also be
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
stressed that the ecosystem was monitored only for a total of four years. As infiltration caused long-lasting chemical changes in forest soil, the treatment could affect tree vitality on a longer term, especially if the infiltration is repeated on the same sites without sufficiently long recovery periods. Acknowledgements The grants provided by the Finnish Funding Agency for Technology and Innovation and the Finnish Ministry of Agriculture and Forestry are gratefully acknowledged. References Bergh JL, Linder S, Lundmark T, Elfving B. The effect of water and nutrient availability on the productivity of Norway spruce in northern and southern Sweden. For Ecol Manag 1999;119:51–62. Bohn HL, McNeal BL, O'Connor GA. Soil Chemistry. 2nd ed. New York: Wiley; 1985. 328pp. Braekke FH. Diagnostiske grensverdier for naeringselementaer i gran- og furunåler. Norwegian, Diagnostic concentrations of nutrient elements in Norway spruce and Scots pine needles, vol. 15/94. Aktuelt fra Skogforsk; 1994. p. 1-11. Cajander AK. Forest types and their significance. Acta For Fenn 1949;56:1-71. Derome J, Lindroos A-J, Helmisaari H-S. Effect of sprinkling infiltration on soil acidity and fertility properties on a forested esker in Central Finland. Recharge systems for protecting and enhancing groundwater resources. Proceedings of the 5th International Symposium on Management of Aquifer Recharge, ISMAR5, Berlin, Germany, 11–16 June 2005, vol. 13. IHP-VI, Series on Groundwater; 2006. p. 264–8. Frycklund C. Total organic carbon retention by filtersand in infiltration pond for artificial groundwater recharge. Aqua Fenn 1995;25:5-14. Helmisaari H-S, Derome J, Kitunen V, Lindroos A-J, Lumme I, Monni S, et al. Sprinkling infiltration in Finland: effects on forest soil, percolation water and vegetation. In:
3371
Peters JH, et al, editor. Artificial Recharge of Groundwater. Rotterdam: Balkema; 1998. p. 243–8. Helmisaari H-S, Illmer K, Hatva T, Lindroos A-J, Miettinen I, Pääkkönen J, et al. Tekopohjaveden muodostaminen: imeytystekniikka, maaperäprosessit ja veden laatu. Metsäntutk l Tied 2003;902:1-209 (In Finnish). Holmes RL. Computer assisted quality control in tree-ring dating and measuring. TreeRing Bull 1983;43:69–78. Kukkola M, Saramäki J. Growth response in repeatedly fertilized pine and spruce stands on mineral soils. Comm Inst For Fenn 1983;114:1-55. Linder S. Responses to water and nutrition in coniferous ecosystems. In: Schultze ED, Zwöjfer H, editors. Potentials and Limitations of Ecosystem Analysis. Ecol StudBerlin: Springer; 1987. p. 180–202. Lindroos A-J, Derome J, Paavolainen L, Helmisaari H-S. The effect of lake-water infiltration on the acidity and base cation status of forest soil. Water Air Soil Pollut 2001;131(1–4):153–67. Lindroos A-J, Kitunen V, Derome J, Helmisaari H-S. Changes in dissolved organic carbon during artificial recharge of groundwater in a forested esker in Southern Finland. Water Res 2002;36:4951–8. Paavolainen L, Kitunen V, Smolander A. Inhibition of nitrification by monoterpenes. Plant Soil 1998;205:147–54. Paavolainen L, Fox M, Smolander A. Nitrification and denitrification in forest soil subjected to sprinkling infiltration. Soil Biol Biochem 2000a;32:669–78. Paavolainen L, Smolander A, Lindroos A-J, Derome J, Helmisaari H-S. Nitrogen transformations and losses in forest soil subjected to sprinkling infiltration. J Environ Qual 2000b;29:1069–74. Smolander A, Kukkola M, Helmisaari H-S, Mäkipää R, Mälkönen E. Functioning of forest ecosystems under nitrogen loading. In: Mälkönen E, editor. Forest condition in a changing environment — the Finnish case. Forestry SciencesDordrecht: Kluwer; 2000. p. 229–47. Smolander A, Ketola R, Kotiaho T, Kanerva S, Suominen K, Kitunen V. Volatile monoterpenes in soil atmosphere under birch and conifers: effects on soil N transformations. Soil Biol Biochem 2006;38:3436–42.
Contents lists available at ScienceDirect
Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v
Artificial recharge of groundwater through sprinkling infiltration: Impacts on forest soil and the nutrient status and growth of Scots pine Pekka Nöjd a,⁎, Antti-Jussi Lindroos a, Aino Smolander a, John Derome b, Ilari Lumme a, Heljä-Sisko Helmisaari a a b
Finnish Forest Research Institute, Vantaa Research Unit, P.O. Box 18, FI-01301 Vantaa, Finland Finnish Forest Research Institute, Rovaniemi Research Unit, P.O. Box 16, FI-96301 Rovaniemi, Finland
a r t i c l e
i n f o
Article history: Received 31 October 2008 Received in revised form 27 January 2009 Accepted 28 January 2009 Available online 9 March 2009 Keywords: Artificial recharge Forest soil Groundwater N transformations Needle nutrients Radial growth Pinus sylvestris (L.)
a b s t r a c t We studied the chemical changes in forest soil and the effects on Scots pine trees caused by continuous sprinkling infiltration over a period of two years, followed by a recovery period of two years. Infiltration increased the water input onto the forest soil by a factor of approximately 1000. After one year of infiltration, the pH of the organic layer had risen from about 4.0 to 6.7. The NH4-N concentration in the organic layer increased, most probably due to the NH4 ions in the infiltration water, as the net N mineralization rate did not increase. Sprinkling infiltration initiated nitrification in the mineral soil. Macronutrient concentrations generally increased in the organic layer and mineral soil. An exception, however, was the concentration of extractable phosphorus, which decreased strongly during the infiltration period and did not show a recovery within two years. The NO3-N and K concentrations had reverted back to their initial level during the two-year recovery period, while the concentrations of Ca, Mg and NH4-N were still elevated. Nutrient concentrations in the pine needles increased on the infiltrated plots. However, the needle P concentration increased, despite the decrease in plant-available P in the soil. Despite the increase in the nutrient status, there were some visible signs of chlorosis in the current-year needles after two years of infiltration. The radial growth of the pines more than doubled on the infiltrated plots, which suggests that the very large increase in the water input onto the forest floor had no adverse effect on the functioning of the trees. However, a monitoring period of four years is not sufficient for detecting potential long term detrimental effects on forest trees. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Shortage of good quality drinking water is an increasing problem in many parts of the world and techniques for producing safe, pure water are therefore highly valuable. In many EU countries artificial recharge of groundwater (AR) is an important means for reaching the goals of the Water Framework Directive. In Finland, 60% of the household water distributed by the waterworks is either natural or artificially recharged groundwater produced by basin and sprinkling infiltration. The proportion of AR groundwater out of total household water production, which is currently ca. 15%, is expected to increase during the next decade. Tampere and Turku, the third and fifth largest cities in Finland, are in the process of adopting the technique. AR groundwater is also widely used in, for example, Sweden. During the production of AR, the concentrations of particulate and dissolved organic carbon in the surface water are reduced by filtration through the overlying soil (e.g. Frycklund, 1995; Lindroos et al., 2002). Chemical and biological processes in the soil also contribute to the reduction in organic carbon. ⁎ Corresponding author. Tel.: +358 10 211 2331; fax: +358 10 211 2203. E-mail address: pekka.nojd@metla.fi (P. Nöjd). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.01.062
Most of the Finnish AR groundwater plants are situated on forested eskers formed at the end of the last ice age. Basin infiltration has been used for decades in several cities for producing AR. Recently, however, many of the eskers have been designated as protected areas or their use has been restricted. Therefore, as the excavation of basins destroys the natural vegetation and topsoil, the establishment of new basins in esker areas is faced with conflicting interests. Since the mid-1990's, Finnish cities have also used a new method, sprinkling infiltration, for AR purposes. In the method, lake or river water is sprinkled directly onto the forest floor from a network of pipes without removing the ground vegetation or the top soil (Helmisaari et al., 1998) (Fig. 1). As no basins are needed, sprinkling infiltration reduces the direct physical disturbance of the esker ecosystems. Sprinkling infiltration has been in use at Jyväskylä, a city with a population of 85 000, since the late 1990s. The artificial recharge of groundwater for supplying the water requirements of large cities involves huge quantities of infiltration water. As the sprinkling infiltration areas are relatively small, the input of water per unit ground surface may be thousands of times higher than the annual precipitation. In addition, the chemical composition of the water used for infiltration differs from that of precipitation. The water source used for producing artificial groundwater for Jyväskylä, Lake Kuusvesi, is a
3366
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
Fig. 1. The principle of producing artificially recharged groundwater (AR) by sprinkling infiltration.
fairly large (22 km2) unpolluted, oligotrophic lake. Its water has a higher content of total organic carbon (TOC), 6.7 mg/l, and a higher mean pH (pH 7.0) than precipitation. It is therefore important to monitor and study the changes in forested esker ecosystems caused by infiltration, and to investigate the possibilities of minimizing the potential adverse effects of such disturbances. The extremely large amounts of water added continuously during sprinkling infiltration have been shown to alter the fluxes of water and nutrients in the soil, resulting in an increase in the pH and base cation concentrations of the uppermost soil layers (Lindroos et al., 2001), and changes in nitrogen cycling (Paavolainen et al., 2000b) and the species abundance of the understorey vegetation (Helmisaari et al., 1993). The results from a study on a fertile forest site in southern Finland (Helmisaari et al., 1998) resulted in recommendations that each site should be infiltrated for a maximum of two growing seasons, followed by a break without irrigation lasting for at least one growing season. In 1998, we initiated a research project at Vuonteenharju, Central Finland, where the water works of the city of Jyväskylä had installed a sprinkling infiltration plant on a nutrient-poor forest site. We monitored the stability of the ecosystem on the forested esker during and after a two-year infiltration period. In this article we report the effects of sprinkling infiltration on soil chemistry, soil nitrogen transformations, and the nutrient status and radial growth of trees. The recovery of the ecosystem was also evaluated. 2. Materials and methods 2.1. The study site The studied infiltration area at Vuonteenharju, Central Finland, is located on an esker 18 km north-east of Jyväskylä. Since 2000, 70% of the household water for the city's population of 85 000 has been produced by artificial recharging through sprinkling infiltration at Vuonteenharju. The infiltration area is located in a Scots pine (Pinus sylvestris) stand on a sub-xeric site representing the Vaccinium vitis-idaea forest site type (Cajander, 1949). The soil consists of relatively coarse, stratified sand deposits. The soil type is humic podzol. Four experimental plots were established in 1998, each 30 × 30 m in size. Two of the plots were infiltrated during 20.9.1999−19.12.2001, and two remained as untreated controls. In addition, another area within the same forest stand, which was infiltrated during 16.11.2002−2.5.2005, was used for studying the effect of infiltration on the diameter growth of Scots pine. The forest had been regenerated during the 1890s. The stand was fully stocked (770 stems/ha, cubic volume 294 m3/ha, dominant height 22.3 m), and had not been treated to silvicultural thinnings for
several decades. The dominant height of 22.3 m at the age of more than 100 years reflects medium-to-low productivity. 2.2. The infiltration treatment During sprinkling infiltration, the annual amount of infiltrated water was c. 600 m3 m− 2, which is about 1000 times the water input from precipitation in this area (Helmisaari et al., 2003). Although the median nitrogen concentrations (NH4-N b0.002 mg/l, NO3-N 0.07 mg/l and Ntot 0.35 mg/l) of the infiltrated lake water were relatively low, the annual amounts of nitrogen added to the forest floor were considerable due to the very large amounts of infiltrated water. A rough estimate of the input of Ntot into the forest floor exceeds 2000 kg/ha/yr and the input of NO3-N was N400 kg/ha/yr, while the magnitude for NH4-N was estimated to be several kilograms/ha/yr. 2.3. Soil samples In order to study the effects of infiltration on soil nitrogen transformations (net N mineralization and nitrification), as well as on factors related to these processes, 18 soil cores (diameter 25 mm) were taken from the organic layer before the start of the treatment (September 4, 1998) and one month (October 25, 1999), one year (September 25, 2000) and two years (September 18, 2001) after the infiltration started. The soil cores were taken systematically at a distance of one meter from the sprinkling infiltration pipes, and bulked to give one composite sample per plot. This means that all the sampling points were subjected to the effects of irrigation. In 2001 samples were also taken from the uppermost 5 cm of the mineral soil layer. The control plots were sampled in a similar fashion. The fresh soil samples were sieved (mesh size 2.8 mm). Soil pH was measured in a water suspension, and the concentrations of NH4-N and NO3-N after KCl extraction as described in Paavolainen et al. (2000b). The rates of net N mineralization and net nitrification were measured in incubation experiments at constant temperature and moisture (Paavolainen et al., 2000a). In order to study changes in the nutrient status (extractable NH4, NO3 and P and exchangeable Ca, Mg and K concentrations) of the topsoil, as well as possible recovery in the nutrient status after infiltration had ceased, soil samples were taken in 2000, 2001, 2002 and 2003 from both the infiltrated and control plots. The organic layer and mineral soil layers (0–5, 5–10, 10–20 and 20–40 cm depth) were sampled at 20 systematically located points on each plot. All the points sampled on the plots were subjected to the effects of irrigation. The samples were combined to give two samples per layer for each plot, and dried at 60 °C. The organic layer samples were milled to pass through a 1 mm sieve, and the mineral soil samples passed through a 2 mm sieve. The extractable NH4 and NO3 concentrations were
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3367
Table 1 pH, concentrations of NH4-N and NO3-N and rates of net N mineralization and net nitrification in the humus layer of the infiltrated and control plots.
pH
NH4-N, mg/kg o.m.
NO3-N, mg/kg o.m.
Net N mineralization (NH4 + NO3)-N, mg/kg o.m./6 wk
Net nitrification NO3-N, mg/kg o.m./6 wk
1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001 1998 1999 2000 2001
Control
Infiltration
3.9 4.0 3.7 3.7 113.8 2.3 17.5 10.9 0.0 0.0 0.6 0.0 − 101.3 5.2 21.1 − 5.3 0.0 0.2 − 0.1 0.0
3.8 5.6 6.7 6.7 112.4 12.1 170.6 117.3 0.0 0.0 11.9 5.5 − 103.4 0.5 − 63.2 − 96.6 0.0 0.3 2.6 − 2.6
Mean for the two replicate study plots. o.m. = organic matter.
Fig. 2. The difference between the infiltrated and control plots in NO3-N concentration (mg/kg of dry matter) one year (2002) and two years (2003) after the termination of infiltration in (a) the organic and (b) the uppermost mineral soil layers.
each sample tree in September 2006. The ring widths were measured to an accuracy of 0.01 mm, and cross-dated using COFECHA software (Holmes, 1983). 3. Results 3.1. Effects of infiltration on soil nitrogen transformations
determined after extraction with 1 M K2SO4 (2 g organic layer or mineral soil/50 ml extractant) by flow injection analysis (FIA). Exchangeable Ca, Mg, K and extractable P concentrations were determined after extraction with 0.1 M BaCl2 (3.75 g organic layer or 15 g mineral soil/150 ml extractant) by inductively coupled atomic emission spectrometry (ICP/AES). 2.4. Needle samples Needle samples were taken from the pines growing close to the infiltration pipes before the treatment (1.10.1998) and after two years of infiltration (18.9.2001), and again from the same trees two years after termination of the treatment (22.9.2003). The radial growth of five of the eight trees sampled per plot was also measured. Two branches from the upper third of the canopy were sampled on each tree. The samples were stored frozen for up to 20 days, the needle age classes separated, and the needles then dried (+50 °C, 72 h) and milled. Total nitrogen was determined on a LECO CHN analyzer, and the P, K, Ca, Mg, Mn, Mo, Zn, Fe, Cu, Al, Ni and B concentrations by ICP/AES after dry digestion (550 °C for 2 h) and extraction with HCl.
The pH of the organic layer was initially slightly below 4 (Table 1), but after one year's infiltration it had risen to 6.7. After both one and two years of continuous sprinkling infiltration, the NH4-N concentration in the organic layer increased (Table 1). The NO3-N concentrations were initially negligible, as well as after one month of sprinkling infiltration. Towards the end of the infiltration period there was a clear increase in the NO3-N concentrations. Net N mineralization was negative before the onset of sprinkling infiltration, indicating either a low rate of gross N mineralization or very high immobilization of nitrogen. During the course of sprinkling infiltration there was an apparent decrease in net N mineralization. Net nitrification was negligible before the onset of sprinkling infiltration, and there was no consistent effect during sprinkling infiltration. In samples taken from the uppermost mineral soil (in 2001) after two years of continuous sprinkling infiltration, net nitrification was negligible on the control plots, but intensive on the sprinkling infiltration plots (Table 2).
Table 3 The Ca, Mg, K, P, NH4-N and NO3-N concentrations (mg/kg d.m.) in the organic layer and mineral soil at depths of 0–5, 5–10, 10–20 and 20–40 cm.
2.5. Growth measurements A total of 22 pines were cored in autumn of 2006 in order to study the effect of infiltration on radial growth. Ten trees were cored on the infiltrated plots (infiltration between 20.9.1999−19.12.2001). Another seven trees were selected from the area infiltrated during 16.11.2002−2.5.2005. Five pines outside the sprinkling infiltration areas, but within 50 m of these two areas, were cored as control trees. An increment core from the bark surface to the pith was taken from
Soil layer
Treatment
Ca
Mg
K
P
NH4-N
NO3-N
Organic layer
Control 2002 — Infiltr. 2003 — Infiltr.
2516⁎ 8033⁎ 7605⁎
308⁎ 1030⁎ 2198⁎
694⁎ 625⁎ 506⁎
143⁎ 53⁎ 55⁎
34.10 51.50 80.70
0.00 0.66 0.31
Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr. Control 2002 — Infiltr. 2003 — Infiltr.
67.5 311.0 402.0 22.4 126.0 173.0 15.0 59.6 62.3 9.6 31.1 42.4
10.70 45.30 53.70 4.20 18.00 22.80 3.09 9.22 10.17 2.12 5.23 7.50
24.90 21.30 24.70 15.20 12.00 16.90 12.50 6.81 7.95 6.10 6.83 9.85
12.20 0.85 0.73 11.70 0.40 1.90 10.50 0.51 1.03 8.20 0.51 2.40
1.84 2.53 2.22 1.84 2.60 2.24 1.84 2.32 1.38 1.84 1.86 1.04
0.00 0.14 0.00 0.00 0.20 0.00 0.00 0.32 0.00 0.00 0.39 0.00
Mineral soil 0–5 cm
5–10 cm
Table 2 pH, concentration of NO3-N and rate of net nitrification in the uppermost mineral soil in 2001.
pH NO3-N, mg/kg d.m. Net nitrification, mg/kg d.m./6 wk Mean for the two replicate study plots. d.m. = dry matter.
Control
Infiltration
4.5 0.0 0.0
6.3 0.5 11.4
10–20 cm
20–40 cm
Control = mean for the control plots without infiltration, 2002 = mean for the infiltrated plots one year after the cessation of the infiltration, 2003 = mean for the infiltrated plots two years after the termination of the infiltration. Values marked with ⁎ are from Derome et al. (2006).
3368
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3.2. Ammonium, nitrate, phosphorus and base cation concentrations in the soil after the termination of infiltration
Fig. 3. The difference between the infiltrated and control plots in Ca concentration (mg/kg of dry matter) one year (2002) and two years (2003) after the termination of infiltration in (a) the organic and (b) the uppermost mineral soil layers.
The NO3-N (Fig. 2) and NH4-N concentrations in both the organic and mineral soil layers on the infiltrated plots were still higher than on the controls one year (2002) after the treatment had ceased, and the NH4-N concentration in the organic layer was even higher after two years (Table 3). The Ca (Fig. 3) and Mg concentrations in both the organic and mineral soil layers on the infiltrated plots were still higher than on the control two years after the termination of infiltration (in 2003) (Table 3). Neither infiltration nor the termination of infiltration had any effect on the K concentrations. Infiltration caused a strong decrease in the extractable P concentrations, and no clear recovery was observed within two years after the termination of infiltration (Table 3).
Fig. 4. Nutrient concentrations in current year (C) and one-year-old (C + 1) needles of Scots pine sample trees on the control and infiltrated plots before infiltration in 1998, after two years of infiltration in 2001, and two years after the termination of infiltration in 2003 (mean and s.d. of 16 trees per treatment).
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
3.3. Tree nutrient status Before the start of infiltration there were only slight differences in needle nutrient concentrations between the control and infiltrated plots, but two years of sprinkling infiltration caused significant changes in the needle nutrient status. Especially the concentrations of N, but also of P, K and B, had increased in the needles on the infiltrated plots by 2001 (Fig. 4). Two years (in 2003) after the end of infiltration, the N, P, K and B concentrations in current-year needles (formed in 2003) had returned to the level prior to infiltration. In contrast, the Ca and Mg concentrations reached their highest levels two years after the termination of infiltration. The nutrient status of the one-year-old needles (formed in 2002) closely followed the patterns in the currentyear needles (Fig. 4). Micronutrient concentrations in the needles (results not shown) did not indicate any signs of nutrient deficiencies during and after infiltration. 3.4. Radial growth Prior to the onset of infiltration, the growth of the sample trees from the two infiltrated areas and the control followed a very similar pattern (Fig. 5a). In 1981, an outbreak of Neodiprion sertifer, an insect that feeds on pine needles, caused a rapid reduction of growth, followed by a gradual recovery during the following 7−8 years. After the onset of infiltration, the mean ring width increased from 0.57 mm in 1999 to 1.47 mm in 2002 (Fig. 5b) on the plots infiltrated
Fig. 5. The mean annual radial growth of Scots pines in the area infiltrated during 20.9.1999–19.12.2001 (red line), in the area infiltrated between 16.11.2002–2.5.2005 (blue line) and in the control (yellow line). The periods shown are 1910–2006 in 5a and 1990–2006 in 5b (infiltration periods shown with thicker lines).
3369
during 20.9.1999−19.12.2001. In 2006 the mean ring width decreased to 0.65 mm, close to the level prior to infiltration. On the area infiltrated between 16.11.2002 and 2.5.2005, there was a strong increase in ring-width in 2004. In 2006 the growth was still higher than that prior to onset of the treatment. The ring widths of the control trees varied only slightly during the period 1999−2006. The year 2000 was an exception, presumably due to favourable weather conditions during the growing season in Central Finland. 4. Discussion We studied the chemical changes on forest soils and the effects on forest trees caused by continuous sprinkling infiltration for a period of more than two years. The experiment consisted of two infiltrated 30 × 30 m plots and two controls of similar size. The experimental design is unsuitable for statistical comparison between treatments due to the lack of sufficient number of replications. A relatively large plot size was considered necessary for preventing surface flow from infiltrated plots to control plots. Given the large plot size, the size of the forest stand and local topography restricted the number of sufficiently homogenous study plots. In addition, as the treatment involved continuous infiltration over a period of more than two years, there would have been technical problems in launching and managing a more complex study design — a side effect of studying the effects of an operating recharge plant. In boreal forest ecosystems, the availability of nitrogen is a key factor limiting tree growth (e.g. Kukkola and Saramäki, 1983; Smolander et al., 2000). This is especially the case on nutrient-poor sites, such as the forested esker at Vuonteenharju used for producing household water by the artificial recharge of groundwater. However, while additional nitrogen is generally beneficial for trees, the excessive total amount of N in the infiltration water could alter the nitrogen cycle (Paavolainen et al., 2000b), resulting in the formation of NO3-N, and thus have highly detrimental effects on the quality of the groundwater. The NO3-N concentrations in percolation water on a fertile site in southern Finland after the termination of infiltration were N10 mg/l (Paavolainen et al., 2000b). We measured both the concentration of mineral N and the rate of net N mineralization, which depicts the amount of easily mineralizable N. The NH4-N concentration in the organic layer of the infiltrated plots increased strongly, most probably due to the NH4 in the sprinkling infiltration water because net N mineralization did not increase (Table 1). Sprinkling infiltration initiated net nitrification in the uppermost mineral soil layers (Table 2), but there was no consistent effect in the organic layer. Both NH4 availability and soil pH regulate nitrification (Smolander et al., 2000). Sprinkling infiltration provides both extra NH4 and a pH that is more optimal for nitrification. In undisturbed, acidic boreal forest soils net nitrification is usually negligible (Smolander et al., 2000). Sprinkling infiltration may increase net nitrification, but the response is apparently dependent on the site characteristics. In a study carried out on a highly fertile forest site in southern Finland, sprinkling infiltration initiated intensive net nitrification in the organic layer soon after the onset of infiltration and also continued at a high rate (Paavolainen et al., 2000b). The main reason for this was the increase in pH (Paavolainen et al., 2000a). At the less fertile site in our study, the low net nitrification in the organic layer was either due to low gross nitrification or intensive immobilization of NO3-N. A lack of substrate could not account for this phenomenon because there was an excess of NH4-N. Interestingly, the conditions for nitrification were apparently more optimal in the mineral soil. One possible reason for the lower net nitrification in the organic layer than in the mineral soil could be higher immobilization of NO3-N in the organic layer. It is also possible that the organic layer contained compounds that inhibited
3370
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
gross nitrification. For instance, volatile monoterpenes have been shown to decrease both the net N mineralization and net nitrification in the organic layer of boreal forests (Paavolainen et al., 1998; Smolander et al., 2006). The elevated NO3-N concentrations in the organic and mineral soil layers lasted for only a relatively short period of time. After a two-year recovery period the differences between the treatment and control were small. This is probably due to the fact that the trees and vegetation effectively utilized the nitrogen input: radial growth of Scots pine more than doubled on the infiltrated areas. However, the NH4-N concentrations were still elevated in the organic layer and the uppermost mineral soil layers after the two-year recovery period. Infiltration caused a clear increase in the exchangeable Ca and Mg concentrations in the organic and mineral soil layers. A similar effect has been reported previously for more fertile sites (Lindroos et al., 2001). The total input of Ca and Mg in infiltration water was extremely large and these cations are readily retained on cation exchange sites on soil particles (primarily organic) in both the organic and mineral soil layers. Neither the Ca nor Mg concentrations showed a decline during the two-year recovery period. The K concentrations in the soil showed only very small changes both during and after infiltration, probably because K ions cannot compete with Ca and Mg ions for cation exchange sites (Bohn et al., 1985; Lindroos et al., 2001; Derome et al., 2006). Infiltration caused a strong decrease in the extractable P concentrations in both the organic and mineral soil layers, and there was no recovery during the two-year period following the termination of infiltration. The relatively high pH most probably caused the immobilization of P, as reflected in the low extractable P concentrations. If the soil pH remains at a relatively high level for many years after the termination of infiltration, as has been reported for a more fertile site in southern Finland (Helmisaari et al., 2003), where no clear recovery was observed within 5 years, then the extractable P concentrations may remain at a low level for many years in our study area owing to the strong link between pH and extractable P concentrations. In boreal forests P has been shown to limit tree growth on the most fertile sites, where nitrogen deficiency is a less severe constraint (Kukkola and Saramäki, 1983). The elevated needle nutrient concentrations after two years of infiltration reflected the increased availability of nitrogen and base cations in the soil (see also Derome et al., 2006). The strong decrease in the extractable P concentrations in the soil on the infiltrated plots was not reflected in the needle P concentrations, which in fact continued to increase after the termination of infiltration. Within two years, the needle nutrient status had almost returned to the level before infiltration. After two years of infiltration, i.e. in autumn 2001, there were, however, some visible symptoms in the needles indicating possible imbalances between certain nutrients. The current-year needles (C) had chlorotic symptoms, and the needles formed during the previous summer (C + 1) had an unusually light green colour. However, this phenomenon was temporary: current-year needles formed and sampled in 2003, i.e. two years after the termination of infiltration, had a healthy green colour. The chlorosis symptoms may be related to an imbalance between nitrogen and magnesium. Magnesium, which is an important constituent of chlorophyll, was the only nutrient that showed a delayed response to infiltration. In 2001, after two years of infiltration, the Mg/N ratios were 0.046 and 0.044 in the C and C + 1 needles, respectively. In 1999 and 2003 they varied between 0.066 and 0.073 in the same trees. Braekke (1994) considered a Mg/N ratio of 0.04 in current-year needles to be a critical value. We conclude that even though the changes in needle nutrient concentration were mostly positive, some detrimental changes (e.g. needle discoloration) occurred. They, however, disappeared rapidly after the infiltration ceased. Sprinkling infiltration doubled the radial growth of the pines at Vuonteenharju. Based on our results, it is not possible to distinguish
between the growth-enhancing effects of an increased water supply and the increase in the nutrient supply to the trees. Some indications are given by Linder (1987), who studied the effects of irrigation and fertilization on the growth of Scots pine in Central Sweden (60° 49' N) during 1974–1984. After 11 years of treatment, the plots which had been treated with both irrigation and fertilization had the highest basal areas. Compared to the combination of irrigation and fertilization, the basal area on the fertilized plots was 84%, on the irrigated plots 49% and on the control plots 38%. Thus, while irrigation produced a slight response, the growth-enhancing effect of nitrogen fertilization was clearly more pronounced. Bergh et al. (1999) obtained volume increments up to four times higher than on the control plots through optimized fertilization and irrigation in two Norway spruce stands in Sweden. At their northern study site (64° 07 'N, 310−320 m a.s.l.), irrigation alone had no effect on volume growth. At the southern experimental site (57° 08 'N, 225−250 m a.s.l.), irrigation also had no significant effect on growth during the first 4-year period. Thereafter, however, the combination of fertilization and irrigation produced a 50% higher volume growth than the fertilized plots without irrigation. In both of the studies, fertilization thus induced a considerably larger growth responses than irrigation. At the northern site of Bergh et al. (1999), representing a slightly higher latitude than that of our study site (62° 20 'N), no irrigation effect was observed. Thus, the large increase in radial growth observed at Vuonteenharju can most probably be attributed to the improved nitrogen supply, which was also reflected in the elevated needle nitrogen concentrations. However, the amounts of infiltration water applied at Vuonteenharju were extremely high compared to those added by Linder (1987) and Bergh et al. (1999). Increased tree growth suggests that the extremely large input of water through the topsoil did not damage the functioning of the trees. This is further supported by the improved supply of the most important macronutrients for trees, as shown by the needle chemistry results. However, there were also observations indicating that caution is required. The discoloration of needles, faint indications of nutrient imbalances in needles, decreased extractable P concentration in the soil and severely altered soil conditions are factors suggesting that tree vitality could be endangered on the long term; especially, if the treatment is repeated on the same site without sufficient recovery periods. 5. Conclusions Virtually continuous infiltration over a period of more than two years caused clear changes in the chemical properties of the organic and uppermost mineral soil layers. The changes were mainly beneficial from the point of view of the nutrient supply of the trees. One notable exception, however, was the extractable phosphorus concentration, which decreased during infiltration and did not recover during the first two years after the termination of infiltration. Needle chemistry analyses showed that the concentrations of important tree nutrients had increased on the infiltrated plots, including the phosphorus. The needle analyses showed no clear signs of nutrient imbalances, although the Mg/N ratio of the needles was close to the critical level. The strong enhancement of radial growth also suggests that continuous sprinkling infiltration over a period of two years did not have any detrimental effects on the vitality of the pines. However, some discoloration of current-year needles, which is a symptom of nutrient imbalances, appeared immediately after the two-year infiltration period. Thus, while we found no direct evidence of detrimental effects on pine following the two-year infiltration period, there were signs of potential risks to the forest ecosystem. These features should be taken into account when planning sprinkling infiltration timetables for the production of artificially recharged groundwater. It should also be
P. Nöjd et al. / Science of the Total Environment 407 (2009) 3365–3371
stressed that the ecosystem was monitored only for a total of four years. As infiltration caused long-lasting chemical changes in forest soil, the treatment could affect tree vitality on a longer term, especially if the infiltration is repeated on the same sites without sufficiently long recovery periods. Acknowledgements The grants provided by the Finnish Funding Agency for Technology and Innovation and the Finnish Ministry of Agriculture and Forestry are gratefully acknowledged. References Bergh JL, Linder S, Lundmark T, Elfving B. The effect of water and nutrient availability on the productivity of Norway spruce in northern and southern Sweden. For Ecol Manag 1999;119:51–62. Bohn HL, McNeal BL, O'Connor GA. Soil Chemistry. 2nd ed. New York: Wiley; 1985. 328pp. Braekke FH. Diagnostiske grensverdier for naeringselementaer i gran- og furunåler. Norwegian, Diagnostic concentrations of nutrient elements in Norway spruce and Scots pine needles, vol. 15/94. Aktuelt fra Skogforsk; 1994. p. 1-11. Cajander AK. Forest types and their significance. Acta For Fenn 1949;56:1-71. Derome J, Lindroos A-J, Helmisaari H-S. Effect of sprinkling infiltration on soil acidity and fertility properties on a forested esker in Central Finland. Recharge systems for protecting and enhancing groundwater resources. Proceedings of the 5th International Symposium on Management of Aquifer Recharge, ISMAR5, Berlin, Germany, 11–16 June 2005, vol. 13. IHP-VI, Series on Groundwater; 2006. p. 264–8. Frycklund C. Total organic carbon retention by filtersand in infiltration pond for artificial groundwater recharge. Aqua Fenn 1995;25:5-14. Helmisaari H-S, Derome J, Kitunen V, Lindroos A-J, Lumme I, Monni S, et al. Sprinkling infiltration in Finland: effects on forest soil, percolation water and vegetation. In:
3371
Peters JH, et al, editor. Artificial Recharge of Groundwater. Rotterdam: Balkema; 1998. p. 243–8. Helmisaari H-S, Illmer K, Hatva T, Lindroos A-J, Miettinen I, Pääkkönen J, et al. Tekopohjaveden muodostaminen: imeytystekniikka, maaperäprosessit ja veden laatu. Metsäntutk l Tied 2003;902:1-209 (In Finnish). Holmes RL. Computer assisted quality control in tree-ring dating and measuring. TreeRing Bull 1983;43:69–78. Kukkola M, Saramäki J. Growth response in repeatedly fertilized pine and spruce stands on mineral soils. Comm Inst For Fenn 1983;114:1-55. Linder S. Responses to water and nutrition in coniferous ecosystems. In: Schultze ED, Zwöjfer H, editors. Potentials and Limitations of Ecosystem Analysis. Ecol StudBerlin: Springer; 1987. p. 180–202. Lindroos A-J, Derome J, Paavolainen L, Helmisaari H-S. The effect of lake-water infiltration on the acidity and base cation status of forest soil. Water Air Soil Pollut 2001;131(1–4):153–67. Lindroos A-J, Kitunen V, Derome J, Helmisaari H-S. Changes in dissolved organic carbon during artificial recharge of groundwater in a forested esker in Southern Finland. Water Res 2002;36:4951–8. Paavolainen L, Kitunen V, Smolander A. Inhibition of nitrification by monoterpenes. Plant Soil 1998;205:147–54. Paavolainen L, Fox M, Smolander A. Nitrification and denitrification in forest soil subjected to sprinkling infiltration. Soil Biol Biochem 2000a;32:669–78. Paavolainen L, Smolander A, Lindroos A-J, Derome J, Helmisaari H-S. Nitrogen transformations and losses in forest soil subjected to sprinkling infiltration. J Environ Qual 2000b;29:1069–74. Smolander A, Kukkola M, Helmisaari H-S, Mäkipää R, Mälkönen E. Functioning of forest ecosystems under nitrogen loading. In: Mälkönen E, editor. Forest condition in a changing environment — the Finnish case. Forestry SciencesDordrecht: Kluwer; 2000. p. 229–47. Smolander A, Ketola R, Kotiaho T, Kanerva S, Suominen K, Kitunen V. Volatile monoterpenes in soil atmosphere under birch and conifers: effects on soil N transformations. Soil Biol Biochem 2006;38:3436–42.