Role of stormflow in reducing N retention in a suburban forested watershed, western Japan
Abstract
[1] To evaluate the role of stormflow in reducing N retention in forested watersheds, we investigated the inorganic N budget of a Japanese suburban forested watershed for 4 years where the proportion of direct flow to precipitation is considerably high (mean is 33%; range is 25–42%). Soil net N mineralization and net nitrification were also measured at middle and lower positions of a slope within the watershed to evaluate nitrate (NO3−) pool size. Annual mean N deposition via throughfall plus stemflow was 15.5 kg N ha−1 yr−1 (13.6–17.1 kg N ha−1 yr−1), which exceeded the threshold value to potentially induce N leaching from forested watersheds. Net nitrification at the middle position was comparable with the lower position. This suggests that the NO3− rich area is large, which could be partly caused by comparable soil moisture conditions with the lower position due to rising groundwater levels during storm events. Annual mean N export was 10.4 kg N ha−1 yr−1 (7.8–12.5 kg N ha−1 yr−1), and subsequent N retention was distinctly low 33% (12–53%). Stormflow accounted for more than 80% of total annual N export. Furthermore, N retention was lower (12 and 27%) in years with higher proportions of direct flow than in years with lower proportions (36 and 53%). Therefore, stormflow is a significant factor in reducing N retention in forested watersheds. The effect of stormflow observation on the comparison of N budgets in other watersheds with different climate and/or hydrologic conditions is discussed.
1. Introduction
[2] The response of forest ecosystems to chronic N deposition has been a concern in Japan [Ohrui and Mitchell, 1997; Baba et al., 2001; Nakaji et al., 2001] as well as in North America and Europe [Johnson and Lindberg, 1992; Bredemeier et al., 1998]. During the past decade, elevated atmospheric N deposition has been recognized to lead to N saturation, defined as an excess of inorganic N (NO3− and NH4+) supply over biotic demand, in forest ecosystems [Aber et al., 1989]. NOx emission, one source of N deposition, has remained high over the last two decades in Japan [Ohara et al., 2007] and North America [Akimoto, 2004], while it has decreased in Europe since 1990 [Akimoto, 2004]. Japan has also been receiving long‐range transport of N compounds from Asia, where NOx emissions have been increased by 2.5 times during 1980–1997 [Akimoto, 2004] and 2.8 times during 1980–2003 [Ohara et al., 2007].
[3] In spite of the high levels of N deposition in Japan, N budgets in Japanese forested watersheds have not been fully evaluated [Ohrui and Mitchell, 1997; Baba et al., 2001; Mitchell et al., 1997; Ohte et al., 2001b; Wakamatsu et al., 2001; Tokuchi et al., 2004]. Analysis of the input‐output N budget for a small watershed provides insight into the response of forest ecosystems to N deposition [Armbruster et al., 2003; Campbell et al., 2004; Likens and Bormann, 1995]. On the other hand, considerable research on N budgets in forested watersheds has been conducted [Dow and DeWalle, 1997; Kahl et al., 1999] and summarized for North America [Johnson and Lindberg, 1992; Campbell et al., 2004] and Europe [Dise and Wright, 1995; Bredemeier et al., 1998].
[4] Many Japanese forested watersheds are characterized by steep topography [Ohrui and Mitchell, 1998; Ohte et al., 2001a]. Steep slopes of watersheds result in small saturated areas that contribute to the large NO3− pool size in the near‐stream zone [Ohrui and Mitchell, 1998], while less steep slopes can have a near‐saturated zone that acts as a N sink due to denitrification [Hill, 1996; Ogawa et al., 2006]. N transformation is a major process in the soil N cycle [Davidson et al., 1992; Hart et al., 1994; Stark and Hart, 1997; Persson et al., 2000], and nitrification is the key process that mobilizes N from forested watersheds, resulting in N leaching [Johnson, 1992; Persson et al., 2000; Yoh, 2001; Fenn et al., 2005]. A larger NO3− pool at the lower positions of slopes within a watershed, corresponding to near‐stream zones, has been recognized in Japanese forested watersheds [Ohte et al., 1997; Ohrui and Mitchell, 1998]. This is because it is generally accepted that net nitrification is significantly higher at lower positions than at middle and top positions of a slope [Garten et al., 1994; Hirobe et al., 1998; Tokuchi et al., 2000]. Therefore, N mineralization and nitrification along a slope should be measured for evaluating the heterogeneous distribution of NO3− pool within a watershed in Japanese forested watersheds with steep slopes.
[5] Steep slopes of watersheds also result in increasing N flushing through water discharge [Creed and Band, 1998; Fujimaki et al., 2008] because stormflow efficiently exports a NO3− pool that is heterogeneously distributed within a watershed. For example, the importance of stormflow in leaching nitrate from upper soils is well documented [Michalzik et al., 2001; Campbell et al., 2004; Ocampo et al., 2006; Bhat et al., 2007; Chiwa et al., 2010]. In addition, a NO3− pool heterogeneously distributed within a watershed is removed by heterogeneous water pathways [Allan et al., 1993; Ohrui and Mitchell, 1998; Creed and Band, 1998; Asano et al., 2006]. Thus, the hydrologic flow path is a significant factor regulating stream water chemistry [Mulder et al., 1990; Ohrui and Mitchell, 1999; Ohte et al., 2001a; Hughes et al., 2007]. Therefore, hydrologic characteristics can be important factors regulating N export in forested watersheds. However, the effect of stormflow on N loss and subsequent N retention has not been fully evaluated in Japanese watersheds.
[6] The Ochozu experimental watershed (OEW) is located 15 km west of the Fukuoka metropolitan area, Kyushu district, western Japan (Figure 1). The OEW is characterized by a high proportion of direct flow to precipitation. The average annual proportion of direct flow to annual precipitation (Qd/P) during 2004–2007 at the OEW was 33% and ranged from 25 to 42% (Table 1). Furthermore, our 4 year observation from 2004 to 2007 was characterized by high variability in the proportion of annual direct flow to annual precipitation. Therefore, observation at the OEW and analysis of annual variation of N budgets and subsequent N retention can test the roles of stormflow in N loss and subsequent reduction in N retention in forested watersheds.
| 2004 | 2005 | 2006 | 2007 | Average | |
|---|---|---|---|---|---|
| Temperature, T (°C) | 16.9 | 16.0 | 17.0 | 16.3 | 16.3 |
| Humidity (%) | 68 | 68 | 71 | 72 | 70 |
| Precipitation, P (mm) | 2201 | 1405 | 2408 | 1436 | 1863 |
| Discharge, Q (mm) | 1127 | 628 | 1421 | 593 | 942 |
| Stormflow, Qs (mm) | 909 | 439 | 1191 | 448 | 747 |
| Direct flow, Qd (mm) | 786 | 354 | 1009 | 371 | 629 |
| Proportion of direct flow to annual precipitation, Qd/P | 0.36 | 0.25 | 0.42 | 0.29 | 0.33 |
[7] The objectives of this study were to evaluate the role of stormflow in reducing N retention in a forested watershed. Specifically, we measured atmospheric N deposition at the OEW, a suburban forested watershed, and soil net N mineralization and net nitrification to evaluate NO3− pool size at the OEW. N export and subsequent N retention at the OEW were also calculated, and interannual variation in N export and subsequent N retention with regard to the proportion of annual direct flow to annual precipitation were analyzed.
2. Materials and Methods
2.1. Partioning Base and Direct Flow
[8] Partitioning base and direct flow was conducted by the method proposed by Hewlett and Hibbert [1967]. This method defines direct flow as beginning from the start of the rise of the hydrograph and ending when the falling limb intercepts an extended line with a slope of 0.0055 l s−1 ha−1 h−1. Stormflow includes both direct and base flows during this period.
2.2. Site Description
[9] This study was conducted at the Ochozu experimental watershed (OEW; 33°38′N, 130°32′E, Figure 1), Kyushu district, western Japan with an elevation range from 160 to 300 m asl. The area, mean slope gradient, and length of the mainstream are 9.5 ha, 0.37, and 265 m, respectively. Annual temperature, annual humidity, and annual precipitation during the study years (2004–2007) at the OEW are shown in Table 1. Approximately 70% of annual precipitation occurred during April to September, corresponding to the growing season. Overstory vegetation was dominated (∼50%) by Japanese cypress (Chamaecyparis obtusa) planted along a stream channel up to the middle slope in 1957. Ridges down to the middle slope within the OEW are covered with a natural‐mixed forest of deciduous species, such as Quercus serrata, Clethra barbinervis, and Rhus succedanea, and evergreen trees, such as Machilus thunbergii, Cinnamomum tenuifolium, Neolitsea sericea, and Quercus glauca (T. Enoki, personal communication, 2010). The bedrock consists of serpentinite and chlorite schist, and the soil type is Cambisols [Ide et al., 2009].
2.3. Water Sample Collection and Analysis
[10] Bulk precipitation (BP), throughfall (TF), stemflow (SF), and stream water (SW) were collected for 4 years from January 2004 to December 2007. Bulk precipitation collectors were installed 2 m above ground in open flats of eastern (260 m asl) and western (230 m asl) ridges (Figure 1). A polyethylene funnel with a diameter of 300 mm was used to collect bulk precipitation. Bulk precipitation samples were collected biweekly. Before sampling, bottles and funnels were cleaned with deionized water.
[11] Throughfall and stemflow were collected in a sampling plot (Japanese cypress forest; 200 m asl) established at the center of the watershed (Figure 1). Five throughfall collectors were placed within the plot 50 cm above the ground. A polyethylene funnel with a diameter of 210 mm was used for collecting throughfall. To avoid contamination, e.g., insects and litter, the throughfall sample was filtered with a polyethylene mesh (mesh width = 200 μm). Three stemflow collectors were installed at breast height (1.3 m) on each tree trunk within the plot. Diameters at breast height (dbh) of the selected trees were 20.8, 20.3, and 26.0 cm. Average dbh of Japanese cypress at the OEW in 2006 was 19.0 cm (±4.6 SD) (T. Enoki, personal communication, 2010). The stemflow collector was constructed with a plastic collar attached to the trunk with silicone sealant as described by Sato et al. [2004]. Throughfall and stemflow sampling was conducted at the same time as the collection of bulk precipitation samples. Before sampling, bottles and funnels were cleaned with deionized water.
[12] Stream water sampling was conducted at discrete temporal intervals above a compound weir with triangular and rectangular notches, which was placed at the mouth of the OEW (160 m asl). Sampling was conducted weekly from January 2004 to March 2007 and biweekly from April to December 2007. Stormflow sampling is important in estimating the annual exports of solutes including dissolved inorganic N (DIN) when the proportion of direct flow to annual precipitation is high [Swistock et al., 1997; Chiwa et al., 2010]. Therefore, stormflow sampling with an automatic sampler (ISCO‐6712, ISCO, NE, USA) was also conducted every 1–3 h during a stormflow period, which included the rising and falling limbs of the hydrograph. Stormflow samplings in 2004, 2005, 2006, and 2007 were conducted six, five, four, and five times, respectively. Water levels at the weir were continually measured every 10 min using a hydraulic pressure sensor (OSASI Tech. Inc., PC‐001). Details of stream water samplings during stormflow at the OEW are described by Chiwa et al. [2010].
[13] The samples were transported to the laboratory within approximately 3 h of being measured at the field site. The samples used to determine NO3− and NH4+ concentrations were filtered with a 0.45 μm–membrane filter (GL science, chromatodisc, 25A). NO3− and NH4+ concentrations were determined using an ion chromatograph (Dionex, DX‐120). Reproducibility of NO3− and NH4+ were 1.9% and 1.7%, respectively. The detection limits of NO3− and NH4+ were 0.26 and 0.32 μmol L−1, respectively. DIN was expressed as the sum of NO3− and NH4+.
2.4. Annual Flux Calculations
| Water (mm) | Inorganic N (NO3− + NH4+) (kg N ha−1 yr−1) | NO3−‐N | NH4+‐N | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 | 2005 | 2006 | 2007 | Average | 2004 | 2005 | 2006 | 2007 | Average | 2004 | 2005 | 2006 | 2007 | Average | 2004 | 2005 | 2006 | 2007 | Average | |
| Input (I) | ||||||||||||||||||||
| BP (A) | 2070 | 1306 | 2241 | 1301 | 1729 | 11.6 | 9.9 | 12.6 | 11.0 | 11.3 | 5.5 | 5.2 | 6.3 | 4.7 | 5.4 | 6.1 | 4.6 | 6.3 | 6.3 | 5.8 |
| TF (B) | 1519 | 901 | 1535 | 879 | 1209 | 15.6 | 13.1 | 11.8 | 13.6 | 13.5 | 10.2 | 8.2 | 7.9 | 7.3 | 8.4 | 5.4 | 5.0 | 3.9 | 6.3 | 5.2 |
| SF (C) | 191 | 97 | 271 | 195 | 189 | 1.6 | 1.3 | 1.8 | 2.9 | 1.9 | 1.2 | 1.0 | 1.5 | 1.8 | 1.4 | 0.4 | 0.3 | 0.4 | 1.1 | 0.6 |
| B + C (D) | 1710 | 998 | 1806 | 1074 | 1397 | 17.1 | 14.5 | 13.6 | 16.6 | 15.5 | 11.4 | 9.2 | 9.4 | 9.2 | 9.8 | 5.8 | 5.2 | 4.3 | 7.4 | 5.7 |
| D − A (NTF) | −360 | −308 | −435 | −227 | −333 | 5.5 | 4.6 | 1.0 | 5.6 | 4.2 | 5.8 | 4.0 | 3.1 | 4.5 | 4.4 | −0.3 | 0.6 | −2.0 | 1.1 | −0.2 |
| Output (O) | ||||||||||||||||||||
| SW (E) | 1127 | 628 | 1421 | 591 | 942 | 12.5 | 9.3 | 12.0 | 7.8 | 10.4 | – | – | – | – | – | – | – | – | – | – |
| SWs (F) | 909 | 439 | 1191 | 448 | 747 | 10.4 | 7.8 | 10.0 | 6.4 | 8.7 | – | – | – | – | – | – | – | – | – | – |
| F/E (G) | 0.81 | 0.70 | 0.84 | 0.76 | 0.79 | 0.83 | 0.84 | 0.83 | 0.82 | 0.84 | – | – | – | – | – | – | – | – | – | – |
| Budget | ||||||||||||||||||||
| D–E (H) | 584 | 370 | 385 | 483 | 456 | 4.6 | 5.2 | 1.6 | 8.8 | 5.1 | – | – | – | – | – | – | – | – | – | – |
| (D–E)/D (I) | 0.34 | 0.37 | 0.21 | 0.45 | 0.33 | 0.27 | 0.36 | 0.12 | 0.53 | 0.33 | – | – | – | – | – | – | – | – | – | – |
- a BP, TF, SF, SW, and SWs indicate bulk precipitation, throughfall, stemflow, stream water, and stream water during stormflow periods, respectively. D − A indicates NTF deposition.
2.5. Soil Sample Collection and Analysis
[17] Net N mineralization and net nitrification rates were determined bimonthly from April to November 2006, which roughly coincided with the growing season. Mineralization and nitrification rates were determined in the humic layer (FH layer) using the buried bag method with in situ incubation [Eno, 1960], and the soil tube method with in situ incubation [Raison et al., 1987] was used at depths of 10 cm (7.5–12.5 cm) and 27 cm (25–30 cm). Soil was collected at two sampling sites, the middle slope (MS) and the lower slope (LS) (Japanese cypress forests; 225 m asl at the MS and 185 m at the LS; Figure 1). At each slope, a 1 × 1 m square grid was established on a 5 × 5 m plot, and five grids among the 25 grids were randomly selected within each plot. After the litter layer was removed from the five grids, the humus (FH) layer was collected. Soil samples from the 7.5–12.5 cm and 25–30 cm depths were collected using a liner sampler (Daiki, DIK‐110B, Japan) with a PVC sample liner tube (5 cm diameter, 30 cm depth; Daiki, DIK‐11A‐R1, Japan). These samples were used as subsamples to determine initial mineral N content. A second set of five samples per plot was buried in the field at the same time and same depth from which they were collected. Buried tubes were covered with a plastic lid on the top to avoid leaching from precipitation. After approximately 60 days, the buried samples were retrieved from each sampling slope.
[18] Following collection, soils were transported to the laboratory within 6 h and sieved through a 2 mm mesh sieve to remove coarse fragments. Ten grams of sieved subsamples of initial (day 0) and incubated (approximately day 60) soils were shaken with 100 mL of 2 M KCl for 1 h to extract NH4+‐N and NO3−‐N. The extract solution was analyzed for NH4+‐N and NO3−‐N. NH4+‐N concentration was determined by the indophenol blue method. NO3−‐N concentration was determined spectrophotometrically after cadmium reduction.
[19] Net N mineralization (mg N 100 g−1 60 days−1) was calculated as the difference of soil NO3−‐N + NH4+‐N concentration between the incubated and initial samples. Net nitrification (mg N 100 g−1 60 days−1) was also calculated as the difference of soil NO3−‐N concentration between the incubated and initial samples. All N data is presented on an oven‐dry weight basis. Net mineralization during the growing season (kg N ha−1) was calculated as the sum of net mineralization of the FH, 0–20 cm, and 20–30 cm layers from April to November 2006, assuming that net mineralization from 7.5 to 12.5 cm and 25–30 cm represented that of 0–20 cm and 20–30 cm, respectively. Transformation from a weight to an area basis was conducted using bulk densities determined from every bimonthly soil sampling.
[20] Soil pH, total C, total N, soil water content, and soil temperature were measured at the MS and LS to investigate soil physicochemical properties affecting net mineralization and net nitrification. Additionally, soil pH, total C, total N, and soil water content were measured in the soil collected for initial mineral N content. In total, 20 soil samples (5 plots × 4 sampling periods) were measured for soil pH, total C, total N, and soil water content during each study period, in each sampling plot, and for each layer. Soil pH was measured in H2O and 1M KCl (soil‐solution mixtures of 1:2.5, m/m) using a glass electrode (Horiba, F‐21, Japan). Total C and N were determined by combustion analysis (Yanako, MT‐700, Japan). Water content was measured by the difference in weight between predried and dried samples. Thermal conductivity was used to measure soil temperature for the FH, 10 cm, and 27 cm layers at sites adjacent to the MS and LS soil sampling plots. Data were recorded at 10 min intervals by a data logger (Campbell, CR‐10X).
2.6. Meteorological Parameters
[21] Precipitation and temperature were continuously measured in open flats of eastern (260 m asl) ridges where bulk precipitation was collected. Precipitation was measured with a tipping‐bucket rain gauge (Takeda Instruments, TK‐1, Certified). Temperature was measured by relative humidity and temperature probes covered with a solar radiation shield 2 m above the ground (Vaisala, HMP45A). Data were recorded at 10 min intervals by a data logger (Campbell, CR‐10X).
3. Results and Discussion
3.1. Hydrological Condition
[22] The average annual proportion of direct flow to annual precipitation (Qd/P) from 2004 to 2007 at the OEW was 33% and ranged from 25 to 42% (Table 1). This value is considerably higher than values reported for other forested watersheds with similar size in Japan (14% [Mitsudera et al., 1984]; 12% [Fukushima, 2007]; 5% [Himeno and Isamoto, 2002]) and the USA (12% [Likens and Bormann, 1995]). The proportion of direct flow to gross precipitation began to increase at a low level of gross precipitation (approximately 10 mm) and reached 73% of individual events for 4 years (Figure 2). This value was also higher than other forested sites with similar size in Japan (59% [Tsujimura et al., 2001]) and in the USA (58% [Hewlett and Helvey, 1970]; 51% [Harr, 1977]). These results suggest that the rapid runoff component in response to rainfall is large at the OEW, resulting in a high proportion of direct flow at the OEW.
[23] A high proportion of direct flow at the OEW could be caused by relatively shallow soil on the slopes of Japanese cypress forests at the OEW (64 ± 19 cm (SD) [Ide, 2008]). This reflects a small water storage capacity (7.6 mm) at the OEW as calculated by Okada et al. [2003] based on 53 stormflow events. This results in frequent overland and subsurface flows, which are components of direct flow at the OEW.
3.2. Bulk Precipitation, Throughfall, and Stemflow Deposition of Inorganic N (NO3− + NH4+)
[24] From 2004 to 2007, the mean annual N deposition via bulk deposition at OEW was 11.3 kg N ha−1 yr−1 and ranged to 9.9–12.6 kg N ha−1 yr−1 (Table 2). This value was comparable to that at urban and suburban forested areas in Japan [Ohrui and Mitchell, 1997; Aikawa et al., 2006].
[25] Annual N deposition via throughfall plus stemflow deposition was higher than via bulk precipitation (Table 2). A positive value of NTF is considered to result from rainfall washing off dry nitrogenous pollutant deposits in the forest canopy [Parker, 1990; Fenn and Kiefer, 1999; Chiwa et al., 2003]. Net throughfall (NTF; TF + SF − BP) N deposition accounted for approximately 30% of TF + SF deposition based on a 4 year average, indicating dry N deposition in an important source of atmospheric deposition as well as wet deposition at the OEW. Bulk precipitation often considerably underestimates dry deposition [Lindberg et al., 1986]. Therefore, it is suggested that the use of bulk precipitation is inappropriate for calculating input‐output N budget and subsequently N retention at the OEW.
[26] Annual N deposition via throughfall plus stemflow was estimated to be 15.5 kg N ha−1 yr−1 based on a 4 year average and ranged from 13.6 to 17.1 kg N ha−1 yr−1 (Table 2). This value was similar to urban and suburban forested watersheds in Japan [Ohrui and Mitchell, 1997; Wakamatsu et al., 2001]. Atmospheric N deposition at the OEW is expected to be higher than N deposition via throughfall plus stemflow because it is reported that 20–40% of atmospheric N deposition to forest stands is retained by the canopy [Cadle et al., 1991; Lovett and Lindberg, 1993; Schulze, 1989; Chiwa et al., 2004]. NTF deposition of inorganic N was mostly composed of NO3−, and that of NH4+ was close to zero (Table 2). Furthermore, NTF deposition of NH4+ was well below zero in 2006. Low NH4+ loads in NTF could be attributed to retention by the forest canopy at the OEW because it is reported that inorganic N, especially NH4+, is taken up by the forest canopy [Parker, 1990; Wilson and Tiley, 1998; Harrison et al., 2000].
[27] Atmospheric N deposition at the OEW was high enough to potentially induce N leaching. Dise and Wright [1995] found that below a N deposition threshold of 10 kg N ha−1 yr−1 no significant N leaching occurred from forests. At intermediate levels of 10–25 kg N ha−1 yr−1, N leaching occurred at some European forested sites, and above 25 kg N ha−1 yr−1, significant leaching occurred at all sites. In addition, the threshold was estimated to be 9–13 kg N ha−1 yr−1 for the whole watershed in the northeastern United States [Aber et al., 2003]. Ohrui and Mitchell [1997] also demonstrated the significant nitrate leaching at two Japanese forests that received more than 10 kg N ha−1 yr−1. However, it should be noted that other factors, such as climate, vegetation type, land use history, hydrologic flow path, etc., influence N leaching from a catchment [Campbell et al., 2004].
3.3. Net N Mineralization and Net Nitrification
[28] Net N mineralization rates at both the middle and the lower slopes (<5.4 mg N 100 g−1 60 days−1 and 28–29 kg N ha−1 growing season−1; Figure 3 and 4, respectively) were comparable to other Japanese cypress (Chamaecyparis obtusa) plantation forests (22 kg N ha−1 yr−1 [Murakami et al., 1990], 1.8 mg N 100 g−1 30 days−1 [Inagaki and Yamada, 2002], 30 kg N ha−1 yr−1 [Tokuchi et al., 2002], 28–31 kg N ha−1 yr−1 [Oyanagi et al., 2004]) where net N mineralization was determined using an in situ incubation method. However, net nitrification was higher than net N mineralization at the two positions (Figures 3 and 4). Relatively high soil pH and low C/N ratio were observed at both positions (Table 3). Net nitrification is usually constrained by low soil pH [White and Gosz, 1987; Killham, 1990; Hirobe et al., 1998; Persson et al., 2000] and a high C/N ratio [Hirobe et al., 1998; Persson et al., 2000; Yoh, 2001]. Therefore, one of the primary causes of the predominance of net nitrification over net N mineralization at the OEW could be high soil pH of the serpentine soil [Brooks, 1987; Kitayama et al., 1998; Miller and Cumming, 2000; Kayama et al., 2005] and low C/N ratio due to moderately high soil temperature, which results in efficient decomposition of plant materials [Yoh, 2001].
| Site | Depth | pH (H2O) | pH (KCl) | C (%) | N (%) | C/N | Water Content (%) | Soil Temperature (°C) |
|---|---|---|---|---|---|---|---|---|
| Lower (LS) | FH layeraa
FH indicates humic layer.
|
5.3 (0.6)bb
Number in parentheses describes standard deviation.
|
4.6 (0.4) | 12.9 (5.2) | 0.72 (0.2) | 18 (3.6) | 47.8 (7.7) | 19.2cc
From April 2006 to September 2006.
(5.0) |
| 10 cm | 5.9 (0.3) | 4.9 (0.3) | 2.5 (1.1) | 0.24 (0.1) | 10 (2.2) | 29.9 (10.0) | 17.9 (4.5) | |
| 27 cm | 6.4 (0.3) | 5.3 (0.2) | 1.4 (0.8) | 0.15 (0.1) | 10 (2.9) | 25.3 (7.0) | 17.5 (4.1) | |
| Middle (MS) | FH layer | 5.3 (0.5) | 4.5 (0.6) | 11.9 (5.7) | 0.67 (0.2) | 17 (2.8) | 46.0 (10.2) | 18.5cc
From April 2006 to September 2006.
(4.5) |
| 10 cm | 5.9 (0.4) | 5.0 (0.2) | 3.2 (1.1) | 0.29 (0.1) | 10 (2.3) | 25.3 (4.7) | 17.9 (4.1) | |
| 27 cm | 6.5 (0.4) | 5.3 (0.3) | 1.7 (0.9) | 0.19 (0.1) | 9 (2.6) | 18.9 (5.1) | 17.5 (3.8) |
- a FH indicates humic layer.
- b Number in parentheses describes standard deviation.
- c From April 2006 to September 2006.
[29] From the viewpoint of N export and subsequent N retention, it should be noted that predominance of net nitrification over net N mineralization was observed even at middle positions (Figures 3 and 4). This result suggests that the NO3− rich area is large at the OEW. Generally, relative net nitrification, defined as the proportion of net nitrification to net N mineralization, decreases with increasing elevation [Garten et al., 1994; Hirobe et al., 1998], and the source of NO3− in stream water is restricted to the area of optimal soil physicochemical condition for nitrification, which corresponds to the lower position of the slope within forested watersheds [Ohrui and Mitchell, 1998].
[30] Similarities of soil physicochemical properties likely explain why net nitrification predominated over net N mineralization even at the middle slope. Variations in net nitrification along a slope are caused by various soil physicochemical properties affecting net nitrification [Garten et al., 1994; Hirobe et al., 1998]. In this study, however, soil physicochemical properties, such as soil pH, C/N ratio, water content, and soil temperature at each soil depth at the middle position were almost the same as those at the lower position (Table 3).
[31] Soil water content is an important factor controlling N transformation including nitrification and NO3− immobilization [Ohte et al., 1997; Creed and Band, 1998; Ohrui and Mitchell, 1998; Tokuchi et al., 2000]. Generally, soil water content decreases with increasing elevation [Enoki et al., 1996; Ohte et al., 1997; Ohrui and Mitchell, 1998; Tromp‐van Meerveld and McDonnell, 2006; Kumagai et al., 2007]. However, in this study, comparable soil water content at the middle position with the lower position was observed, which could result in an increased area having optimal moisture conditions for nitrification. Because soil is relatively shallow on the slopes of Japanese cypress forests at the OEW [Ide, 2008], this increase could have been caused by rising groundwater level during storm events. NO3− immobilization is considered to be a key process of net nitrification and subsequently N retention [Stark and Hart, 1997; Hirobe et al., 2003]. Although only the net rate of N transformation was measured in this study, N transformation including NO3− immobilization at the middle position could be similar to that at the lower position, resulting in predominance of net nitrification over net N mineralization at middle positions.
3.4. Stream Water Concentration of Inorganic N (NO3− and NH4+) and Annual Exports
[32] Annual average concentrations of NO3−, the dominant form of inorganic N, in stream water sampled at discrete intervals were 54–69 μmol L−1 from 2004 to 2007 (Table 4). These values were higher than in many other forested watersheds in Japan (5–50 μmol L−1 [Shibata et al., 2001], 4–11 μmol L−1 [Zhang et al., 2008]) and the United States (1–2 μmol L−1 [Stottlemyer and Troendle, 1992], <1 μmol L−1 [Vanderbilt et al., 2003], 5 μmol L−1 [Sickman et al., 2002]) but were similar with urban or suburban forested watersheds in Japan receiving high atmospheric N deposition (≈100 μmol L−1 [Ohrui and Mitchell, 1997], 76 μmol L−1 [Okochi and Igawa, 2001], 80–90 μmol L−1 [Shibata et al., 2001], 59–66 μmol L−1 [Zhang et al., 2008]).
| NO3−‐N | NH4+‐N | |||||||
|---|---|---|---|---|---|---|---|---|
| 2004 | 2005 | 2006 | 2007 | 2004 | 2005 | 2006 | 2007 | |
| Average | 68.9 | 54.4 | 61.4 | 65.9 | 1.3 | 1.6 | 1.4 | 2.7 |
| Median | 64.9 | 49.5 | 57.8 | 53.7 | 0.1 | 0.1 | 0.3 | 1.2 |
| Minimum | 34.9 | 27.3 | 25.5 | 30.0 | ND | ND | ND | ND |
| Maximum | 122.7 | 111.1 | 116.5 | 136.2 | 6.3 | 12.0 | 18.2 | 11.4 |
| SD | 21.5 | 21.5 | 25.2 | 30.9 | 1.8 | 2.9 | 3.1 | 3.7 |
| n | 53 | 53 | 52 | 37 | 53 | 53 | 52 | 37 |
- a Units are in μmol L−1. SD, standard deviations; ND, not detected.
[33] NO3− concentrations remained high during the growing season (Figure 5). Little seasonal variation in NO3− concentrations in stream water observed in many forested watersheds of Japan [Ohrui and Mitchell, 1997; Ohte et al., 2001a] is thought to be due to high N mineralization and nitrification rates during the growing season as well as hydrologic factors [Ohrui and Mitchell, 1997; Mitchell et al., 1997; Ohte et al., 2001a]. This would be true of the OEW. On the other hand, NO3− concentrations in stream water in N‐limited forested watersheds are generally low during the growing season in the northern United States and Europe [Murdoch and Stoddard, 1993; Stoddard, 1994; Mitchell et al., 1996; McHale et al., 2000].
[34] N export at the OEW based on a 4 year average was 10.4 kg N ha−1 yr−1 and ranged from 7.8 to 12.5 kg N ha−1 yr−1 (Table 2). This indicates a substantial N loss. Average N retention of inorganic N was only 33% and ranged from 11 to 53% (Table 2). These values were considerably lower than those for many forested watersheds; for example, 66% at Kiryu in Japan [Ohrui and Mitchell, 1997], 52–70% at Shichinohe in northeastern Japan [Baba et al., 2001], 78% (−19–100%) in 53 forests in Germany [Brumme and Khanna, 2008], 72–73% in the Catskill Mountains of southeastern New York, United States [Lovett et al., 2000], and 69% (27–99%) in 24 forests in the northeastern United States [Campbell et al., 2004].
3.5. Effect of Stormflow on N Loss
[35] Stormflow appears to increase N loss and subsequent lower N retention at the OEW. Stormflow accounted for more than 80% of total annual N export every year (Table 2). In addition, N retention was lower in years with a higher proportion of annual direct flow to annual precipitation (2004 and 2006) than in years with a lower proportion (2005 and 2007, Table 2). Bhat et al. [2007] demonstrated that surface runoff (direct flow) during storm events is an important contributor to N export from forested watersheds. Also, hydrologic flow path is an important factor regulating N exports from forested watersheds [Michalzik et al., 2001; Ohte et al., 2001a; Ocampo et al., 2006].
[36] Lower N retention by stormflow could be caused via the following two processes: (1) biogeochemical processes, including increases in the NO3− rich area, and (2) hydrological processes via efficient export of NO3−. Stormflow increase the size of areas with optimal moisture conditions for nitrification due to rising groundwater levels during storm events, as previously mentioned.
[37] Stormflow includes efficient export of the NO3− pool from the upper soil surface where NO3− is abundant. Direct flow is composed of overland flow and subsurface flow. Frequent overland and subsurface flow at the OEW would efficiently export NO3− from the watershed because flow occurs on or within the upper soil surface. Changes in NO3− concentrations in stream water corresponded well with changes in discharge during both growing and dormant seasons (Figure 5). This suggests that changes in hydrologic condition are important drivers for N export at the OEW.
[38] Effects of stormflow on N export could differ among watersheds with different climate and/or hydrologic conditions. The large NO3− pool size as a result of nitrification during the growing season would be efficiently exported from the OEW because approximately 70% of annual precipitation occurs during the growing season. In Japan, both rates of N nitrification and precipitation are high during the same season, mainly in summer, but trends differ in southeastern United States where rates of N nitrification are high in summer and precipitation is high in winter [Ohte et al., 2001b]. Therefore, observation of stormflow is essential for calculating the input‐output N budget in a forested watershed, especially when comparing other watersheds with different climate and/or hydrologic conditions.
4. Summary and Conclusions
[39] To evaluate the role of stormflow in reducing N retention in forested watersheds, we investigated the input‐output N budget and subsequent N retention for 4 years at the OEW, a suburban forested watershed in western Japan. The OEW is characterized by a high proportion of direct flow to precipitation. Bulk precipitation, throughfall, and stemflow were collected to measure atmospheric N deposition, and stream water samples were collected at discrete intervals and during stormflow to calculate N export from the OEW. Soil net N mineralization and net nitrification were also measured at the middle and lower positions of the slopes within the watershed to evaluate NO3− pool size.
[40] Based on a 4 year average, annual mean N deposition via bulk precipitation and throughfall plus stemflow were 11.3 kg N ha−1 yr−1 (9.9–12.6 kg N ha−1 yr−1) and 15.5 kg N ha−1 yr−1 (13.6–17.1 kg N ha−1 yr−1), respectively. Atmospheric N deposition was comparable to rates in urban or suburban forested watersheds in Japan. These values were above the threshold value that could potentially cause N leaching from a forested watershed.
[41] Although the net N mineralization rate was not extremely high, net nitrification predominated over net N mineralization. The predominance of net nitrification over net N mineralization was probably caused by high soil pH and low C/N ratio at the OEW. Furthermore, this was observed even at the middle slopes, which could be partly caused by comparable soil moisture condition with the lower position due to rising groundwater levels during storm events. Therefore, it is suggested that the NO3− rich area is large at the OEW.
[42] NO3− concentration in stream water was considerably higher throughout the year than in many other forested watersheds. Annual mean N export from stream water was calculated to be 10.4 kg N ha−1 yr−1 based on a 4 year average and ranged from 7.8 to 12.5 kg N ha−1 yr−1. Subsequent N retention was distinctly lower (33%) than in other forested watersheds in most cases.
[43] Stormflow accounted for more than 80% of the total annual N export every year. Also, N retention was lower (12 and 27%) in years with a higher proportion of annual direct flow to annual precipitation than in years with a lower proportion (36 and 53%). In addition, corresponding changes in NO3− concentrations in stream water with those in discharge during both growing and dormant seasons suggest that changes in hydrologic condition are important drivers for N export at the OEW. These results confirmed that stormflow is a significant factor in reducing N retention in forested watersheds via (1) biogeochemical process, including increases in the NO3− rich area, and (2) hydrological process via efficient export of the NO3− pool. In addition, it was demonstrated that observation of stormflow is essential for calculating the input‐output N budget in a forested watershed, especially when comparing other watersheds with different climate and/or hydrologic conditions.
Acknowledgments
[44] We thank T. Enoki of Kyushu University Forest for advice on the manuscript. This work was partially supported by Grants‐in‐Aid for Scientific Research (14360088, 16006809, 17380096, and 1820814) and for Young Scientists (17780125) from the Ministry of Education, Culture and Science, and Technology, Japan and “Rehabilitation of Ariake Bay and demonstration of rehabilitation technologies” in a research and development program for resolving critical issues commissioned by the Ministry of Education, Culture, Sports, Science and Technology, Japan. We also thank the Young Researcher Support Project of Faculty of Agriculture, Kyushu University, for supporting our studies.
