Hydrological and Biogeochemical Controls of the Seasonality of Particulate and Dissolved Nitrogen Exports During Rainfall Events From a Forested Watershed in Monsoon Asia, Japan

Few studies have investigated the effects of hydrological conditions such as rainfall magnitude and soil nitrogen dynamics on nitrogen export from forests during rainfall events. In this study, the seasonal export of particulate and dissolved nitrogen during 24 rainfall events (range: 3.0–417 mm) in a temperate monsoon forest watershed in Japan was measured along with the seasonality of the soil nitrate ( NO3− ${{\text{NO}}_{3}}^{-}$ ) pool size and the nitrification rate. The increase in nitrogen export with increasing streamflow was larger for particulate nitrogen than for NO3− ${{\text{NO}}_{3}}^{-}$ and dissolved organic nitrogen, but NO3− ${{\text{NO}}_{3}}^{-}$ was always the most abundant nitrogen component exported, even during extreme rainfall. For all rainfall events in summer, measured NO3− ${{\text{NO}}_{3}}^{-}$ exports were higher than the averaged NO3− ${{\text{NO}}_{3}}^{-}$ exports, regardless of rainfall magnitude or preceding rainfall. Nitrification activity in the soil was high in summer because of the high temperature and wet conditions of the soil. Nitrification also occurred in the upper slope in summer, whereas during other seasons it is hindered by dry conditions. These results indicate that nitrogen dynamics in the soil also affect NO3− ${{\text{NO}}_{3}}^{-}$ export during rainfall. Comparisons with previous studies suggested that the effects of soil nitrogen dynamics on NO3− ${{\text{NO}}_{3}}^{-}$ concentrations in stream water may be stronger during rainfall than during base flow conditions. They also showed that regions with wet summers due to rainfall are more sensitive to NO3− ${{\text{NO}}_{3}}^{-}$ export, compared with other regions, and that increased summer rainfall can have a particularly large impact on nitrogen export from forest watersheds.


Introduction
Nitrogen exported from forest watersheds is a source of nutrient loss for forest ecosystems but a nutrient supply for downstream aquatic ecosystems.However, excess nitrogen exported from forested watersheds contributes to eutrophication in downstream watersheds, degrades drinking water quality and adds to greenhouse gas emissions through microbial nitrogen transformation.Addressing these issues requires accurate assessment of the amount and process of nitrogen export from forest watersheds.
Previous studies have shown that nitrogen exported from forest watersheds during rainfall events constitutes a large proportion of the annual nitrogen export (Ahearn et al., 2004;Chiwa et al., 2010;Hoover & Mackenzie, 2009;S. Inamdar et al., 2015;Kunimatsu et al., 1999Kunimatsu et al., , 2006)).For example, in their study of the Cosumnes River Watershed in California, USA, where most rainfall occurs between December and March, Ahearn et al. (2004) found that most of the annual nitrate (  NO3 − ) export also occurs during this period.Kunimatsu et al. (2006) conducted a 3-year weekly survey and surveyed five rainfall events in the Aburahi-S forest watershed (3.34 ha); they calculated that the annual runoff of total nitrogen (TN) was 9.95 kg/ha and that  NO3 − constituted 72.7% of TN, of which 73% was exported due to rainfall events.Therefore, although rainfall events are of short duration, ranging from a few hours to several tens of hours, they are a major determinant of the amount of nitrogen exported from forest watersheds.Moreover, as rainfall intensity and patterns have been changing in many areas due to climate change (Katsuyama et al., 2021;Trenberth, 2011;Westra et al., 2014), the impact of rainfall events on the amount of nitrogen exported from forests should be examined.
The processes resulting in dissolved nitrogen (  NO3 − and dissolved organic nitrogen [DON]) export from forest watersheds during rainfall events have been well studied.Temporal variations in both the  NO3 − concentration and source in stream water during rainfall events were shown to be strongly influenced by hydrological processes and nitrogen distribution in the watershed (Burns & Kendall, 2002;Creed & Band, 1998; S. P. Inamdar et al., 2004; Abstract Few studies have investigated the effects of hydrological conditions such as rainfall magnitude and soil nitrogen dynamics on nitrogen export from forests during rainfall events.In this study, the seasonal export of particulate and dissolved nitrogen during 24 rainfall events (range: 3.0-417 mm) in a temperate monsoon forest watershed in Japan was measured along with the seasonality of the soil nitrate (  NO3 − ) pool size and the nitrification rate.The increase in nitrogen export with increasing streamflow was larger for particulate nitrogen than for  NO3 − and dissolved organic nitrogen, but  NO3 − was always the most abundant nitrogen component exported, even during extreme rainfall.For all rainfall events in summer, measured  NO3 − exports were higher than the averaged  NO3 − exports, regardless of rainfall magnitude or preceding rainfall.Nitrification activity in the soil was high in summer because of the high temperature and wet conditions of the soil.Nitrification also occurred in the upper slope in summer, whereas during other seasons it is hindered by dry conditions.These results indicate that nitrogen dynamics in the soil also affect  NO3 − export during rainfall.Comparisons with previous studies suggested that the effects of soil nitrogen dynamics on  NO3 − concentrations in stream water may be stronger during rainfall than during base flow conditions.They also showed that regions with wet summers due to rainfall are more sensitive to  NO3 − export, compared with other regions, and that increased summer rainfall can have a particularly large impact on nitrogen export from forest watersheds.
•  NO3 − was the most abundant nitrogen component exported, even during extreme rainfall events • For summer rainfall events, measured  NO3 − exports were higher than averaged ones • The relationship between  NO3 − production potential in soil and nitrogen export applies to rainfall events as well as baseflow

Supporting Information:
Supporting Information may be found in the online version of this article.S. Inamdar et al., 2015;Muraoka & Hirata, 1988).For instance, Muraoka and Hirata. (1988) found that, in the forest watershed of Mt.Tsukuba, Japan,  NO3 − concentrations in stream water increased after the peak in rainfall, suggesting that it is a source of  NO3 − accumulation in the soil surface layer.S. P. Inamdar et al. (2004) reported that  NO3 − concentrations in stream water during rainfall were highest before the peak of stream water discharge.They also identified  NO3 − accumulated in glacial till groundwater as the source of the  NO3 − detected in a forested watershed in the Adirondack Mountains of New York, USA.

Seasonal variations of
NO3 − concentrations in forested streams during baseflow have been discussed in relation to hydrologic processes and  NO3 − distribution in the watershed, as well as seasonal variations in soil-plant nitrogen dynamics, including  NO3 − uptake by plants and mineralization and nitrification in the soil.For example, in Northeast American forests, where precipitation is higher in winter,  NO3 − concentrations in stream water are higher in winter and lower in summer, a difference attributed to the transfer of  NO3 − from soil to vegetation in summer (during vegetation growth) versus winter (no vegetation growth) (Judd et al., 2007;Likens & Bormann, 1995;Mitchell et al., 1996).However, in Japan, where summer precipitation is more frequent,  NO3 − concentrations in forest stream water are often higher in summer than in winter, indicating that the greater hydrologic  NO3 − transport capacity due to high rainfall is more important than the effect of plant  NO3 − uptake during the summer growing season (Ohte, 2012;Ohte et al., 2001;Ohte & Katsuyama, 2019).Studies focusing on inter-forest comparisons have also shown that the C:N ratio of the organic and soil layers strongly affects both nitrification in the soil during baseflow and  NO3 − export from the forest watershed (Aber et al., 2003;Emmett et al., 1998;Kuroiwa et al., 2011;Lovett et al., 2002).Other studies of forests have shown that the C and N concentrations and the C:N ratios of organic matter and soil layers are strongly influenced by the tree species of the forest (Lovett et al., 2002(Lovett et al., , 2004;;Urakawa et al., 2016).These observations highlight the effects of nitrogen dynamics in soil and vegetation on  NO3 − export processes during baseflow.
In contrast to the above-cited research, the effect of the seasonality of soil-plant nitrogen dynamics on  NO3 − export during rainfall events has not been well-studied.In one of the few examples, a study conducted in the northeastern United States, where there is snowmelt in spring, showed that  NO3 − export relative to streamflow was higher after spring rainfall than after rainfall during other seasons, possibly due to the export of  NO3 − previously retained in the snowpack and less  NO3 − uptake by plants (Vaughan et al., 2017). NO3 − export during rainfall was also lower in autumn, that is, the litterfall season, than during other seasons, which was attributed to the lower nitrification rate, enhanced assimilation and increased denitrification occurring in response to the increase in dissolved organic carbon supply from fresh litter (Sebestyen et al., 2014;Vaughan et al., 2017).These studies point to an important effect of soil-plant nitrogen dynamics on nitrogen export during rainfall.However, for a full understanding of the influence of climate change on forest nitrogen export, knowledge of the links between forest nitrogen dynamics and nitrogen export in response to rainfall events in forests with different climates, vegetation and soil patterns is required.Because the amount of  NO3 − exported from forests is strongly influenced by export during rainfall, clarification of these relationships will provide a better understanding of the process of nitrogen export from forests.
Studies have also shown much larger increases in the export of particulate nitrogen (PN) than dissolved nitrogen with increasing rainfall (Correll et al., 1999;S. Inamdar et al., 2015;Kunimatsu et al., 2006;Wiegner et al., 2009).Therefore, determination of particulate and dissolved nitrogen exports is necessary to comprehensively evaluate the effects of rainfall intensity and patterns on the amount of nitrogen exported from forests.However, few studies have simultaneously examined the magnitude of rainfall and seasonality of soil nitrogen dynamics with respect to dissolved nitrogen and PN exported from forests during rainfall events.
In this study, based on the hypothesis that seasonal variations in nitrogen dynamics in forest soils affect nitrogen export from forest watersheds during rainfall, we examined the effects of rainfall magnitude and the seasonality of soil nitrogen dynamics on the export of dissolved nitrogen and PN from forests during rainfall.We measured the seasonality of soil inorganic nitrogen availability and nitrification rate in a temperate monsoon forest watershed in southern Shiga Prefecture, Japan, as well as the amount of particulate and dissolved nitrogen exported during 24 rainfall events ranging in magnitude from 3.0 to 417 mm among different seasons.Meteorological Agency, 2023).The average annual precipitation during the same period was 1,595 mm (range: 1,104-2,041 mm, standard deviation: 228 mm) (Japan Meteorological Agency, 2023).There is little snowfall in the study area, even in winter, and all 24 events included in this study were rainfall events.The average rainfall during the 5 months of the rainy season (June-October) is 946 mm (Figure 2).In the cypress plantation forest of Ab-S, there is little understory vegetation due to the limited light conditions below the canopy.From   S1).
The average stream channel slope is not steep (11.5°), unlike the surrounding hill slopes, the majority of which have a gradient of 30°-40°.Most of the streambed and streambank is exposed weathered granite.Soil depths at the lower, middle, and upper areas of the slope are in the range of 36-100 cm (average: 61 cm), 70-156 cm (average: 117 cm), and 36-156 cm (average: 82 cm), respectively.N4 values, defined as the number of blows required for a ground penetration of 4 cm, were measured at least four times at each site using a cone penetrometer with a diameter of 9.5 mm, weight of 1.17 kg and drop distance of 20 cm.Sites with N4 values >100 were defined as constitutional bedrock, and soil depth was defined as the distance from the ground surface to the bedrock.The soils in Ab-S are relatively immature; the A 0 horizon layer is ∼0-5 cm and the A horizon layer ∼5-20 cm.
The average annual litterfall from September 2000 to August 2002 was 5,550 kg/ha in the lower cypress plantation forest and 3,650 kg/ha in the upper broadleaved forest (Osaka et al., 2011).Total dissolved nitrogen (TDN) and  NO3 − deposition from July 2012 to June 2013 (374 days, rainfall 1,675 mm) were 5.64 and 3.00 kgN/ha for rainfall and 14.4 and 8.66 kgN/ha for throughfall, respectively (Osaka et al., 2016).A long-term forest monitoring site, the Kiryu Experimental Watershed, is located 26 km east of the study site.At this long-term forest monitoring site, Tokuchi et al., 2013, who measured dissolved inorganic nitrogen deposition from 1990 to 2005, reported no clear temporal trend in the annual atmospheric N input.The deposition of  NO3 − in the urban area of Otsu City, located about 40 km northwest of the study site, showed no clear trend of increase or decrease from the 1990s to the 2010s, although there was an increase or decrease from 1 year to the next (Lake Biwa Environmental Research Institute, 2023).No drastic change in nitrogen deposition was expected during the observation period in this area, which included the study site.

Sample Collection and Field Observation
Stream water was sampled for chemical analysis during a total of 24 rainfall events, from the beginning of the rainfall until the post-rainfall decrease in stream water discharge rate.Stream water was collected manually or using a water sampler (3,700 or 6,712; Teledyne ISCO, Lincoln, NE, USA) approximately once every hour during the rainfall event, directly above a weir located at the end of the watershed.Some of the collected samples were immediately filtered through cellulose acetate membrane filters (0.45-μm pore size).Both filtered and unfiltered samples were stored in polypropylene bottles at 4°C.The stream water discharge rate was calculated from the overflow water level measured at the weir.The amount of rainfall was measured using a tipping bucket and data logger in the open meteorological stations (outside the experimental watershed, ∼200 m from the weir).
Soil samples were collected every 1-3 months on a total of 11 occasions between June 2012 and December 2013.Soil samples were collected at 0-10 and 20-30 cm at the lower and upper parts of three slopes, respectively (Figure 1).The samples were immediately sieved through a 2.0-mm screen and extracted within 1 day using 2N KCl.The extracts were stored in polypropylene bottles at −25°C.Soil samples used in the measurement of net nitrification rates were incubated for 40 days at the in situ temperature and moisture level, with the net nitrification rate calculated as the difference between the initial and post-incubation  NO3 − concentrations.Surface soil temperatures (at depths of 5 and 20 cm) were measured using a thermometer and data logger (model TR-71u: T&D Corporation, Tokyo, Japan) at 30-min intervals.The 1-week average values of the surface soil temperatures before soil collection were used as the in situ temperatures.

Chemical Analysis and Calculation for Nutrient Export
The  NO3 − concentration in stream water samples was determined by ion chromatography (HIC-6A; Shimadzu, Tokyo, Japan; Compact IC 762; Metrohm, Herisau, Switzerland; ICS-1100; Thermo Fisher Scientific, Waltham, MA, USA).TDN and TN (sum of all nitrogen forms) were measured by ultraviolet absorption (UV-2450: Shimadzu) after oxidation of the filtered and unfiltered samples, respectively, with potassium persulfate.Ammonium (  NH4 + ),  NO3 − and dissolved silica (SiO 2 ) were colored by treatment with indophenol, naphthylethylenediamine and molybdenum yellow, respectively, and their absorbance was measured (UV-2450; Shimadzu).The PN concentration was calculated as the difference between TN and TDN.Unfiltered water was collected with a 10-mL volumetric pipette for TN analysis.The pipette tip was about 1 mm in diameter, so the sizes of TN and PN measured in this study were approximately less than 1 mm.DON was calculated as the sum of the nitrogen content of  NO3 − ,  NH4 + , and  NO2 − subtracted from TDN.The  NO3 − concentration in the KCl soil solutions was determined by flow injection analysis (OG-F1-300S; Ogawa & Co., Tokyo, Japan).
Nutrient export from the watershed was calculated from the rainfall event data by the "integration interval-loads" method of Kunimatsu et al. (2006).The hydrograph was separated into direct runoff and baseflow using the method described in Hewlett and Hibbert (1967).The duration of a rainfall event was considered to be identical to the duration of the direct runoff.Some of the event data used in this study were obtained by Kunimatsu et al. (2006; some data for events 7-9) and Osaka et al. (2016; some data for events 12-19).

Streamflow for Rainfall Events at Ab-S
The magnitude of the rainfall for the 24 rainfall events included in this study ranged from 3.0 to 417 mm (Table 1).The 417-mm rainfall event was unusually large and occurred within ∼24 hr with the arrival of typhoon no.11 in 2001.In the Ab-S watershed, streamflow increased with increasing rainfall, although the increase was small after events with <30 mm of rainfall (Figure 3).As shown in Figure 3, there were no marked changes in the relationship between rainfall and direct runoff during the 20-year study period.The mean direct runoff rates for rainfall events of 0-30 mm, 30-100 mm, and >100 mm were 10.4% (range: 1.19%-37.4%),44.4% (range: 17.1%-75.4%),and 46.1% (range: 32.8%-62.6%),respectively.During heavy rainfall events, direct runoff was usually equivalent to ∼50% of the amount of rainfall.

Particulate and Dissolved Nitrogen Exports During Rainfall Events
The amounts of TN, PN, DON, and  NO3 − exported during rainfall events in the Ab-S watershed were 0.007-3.523kgN/ha, 0.001-1.208kgN/ha, 0.001-0.317kgN/ha, and 0.005-1.932kgN/ha, respectively (Table 1).In all rainfall events, most of the exported nitrogen was in the form of  NO3 − ; the average ratio of  NO3 − to TN was 70.0% (range: 44.8%-96.3%).However, the relationship of each exported nitrogen component to streamflow during rainfall was different, as indicated by the power functions in Figure 4.The power exponents of TN, DON, and  NO3 − were 0.97, 0.98, and 0.95, respectively, and they increased almost linearly with increasing streamflow.The power exponent between PN export and streamflow was 1.26, indicating that PN export per unit flow rate increased with increasing streamflow.The ratio of PN export to TN export therefore also increased as the streamflow rate per rainfall event increased (Figure 5).
The weekly sampling data showed slightly higher  NO3 − concentrations from 1999 to 2008 than during other periods (Table S1), attributable to the slight thinning that took place in September-October 1999.However, analysis of covariance with the logarithm of exported  NO3 − as the dependent variable, logarithm of streamflow as the covariate, and slight thinning (rainfall events 6-13) or not (remaining events) as a fixed factor, showed that the effect of the fixed factor was not significant (p = 0.29).Thus, when considered in isolation, the slight thinning that took place at the study site did not have a significant effect on  NO3 − export during rainfall events.slopes, with median  NO3 − concentrations of 1.06 mgN/kg dry soil (0-10 cm) 0.66 mgN/kg dry soil (20-30 cm) at the lower slopes and 0.17 mgN/kg dry soil (0-10 cm) and 0.14 mgN/kg dry soil (20-30 cm) at the upper slopes.The median net nitrification rates were 0.26 mgN/kg dry soil/day (0-10 cm) and 0.09 mgN/kg dry soil/ day (20-30 cm) at the lower slopes and 0.01 mgN/kg dry soil/day (0-10 cm) and 0.02 mgN/kg dry soil/day (20-30 cm) at the upper slopes.Soil nitrogen concentrations were higher at 0-10 cm than at 20-30 cm at both the lower and upper slopes, with median soil nitrogen concentrations of 0.36% (0-10 cm) and 0.17% (20-30 cm) at the lower slopes and 0.31% (0-10 cm) and 0.11% (20-30 cm) at the upper slopes.Soil C:N ratios were slightly higher at the upper slopes than at the lower slopes.The median C:N ratios were 17.1 (0-10 cm) and 16.0 (20-30 cm) at the lower slopes and 19.0 (0-10 cm) and 18.3 (20-30 cm) at the upper slopes.The weighted water content was higher at the lower slopes than at the upper slopes, with median values of 36.8% (0-10 cm) and 29.0% (20-30 cm) at the lower slopes and 20.9% (0-10 cm) and 16.5% (20-30 cm) at the upper slopes.

Soil Nitrogen Concentration and Net Nitrification Rate
Figure 7 shows the relationships of soil temperature with the nitrification rate and  NO3 − concentration in soil; Figure 8 shows the relationship between the weighted water content of soil and the net nitrification rate.In summer, when the soil temperature was high, the high nitrification rate resulted in the rapid production of 10.1029/2023WR034756 8 of 15  NO3 − in soil (Figure 7).In addition to soil temperature, the nitrification rate was affected by the soil water content, which was higher at sites with a high weighted water content that coincided with a high soil temperature (Figure 8).

Contributions of Dissolved and Particulate Nitrogen Exports to Total Nitrogen Export
In this study, the increase in exported nitrogen with increasing streamflow was greater for PN than for either  NO3 − or DON, resulting in an increase in the contribution of exported PN to exported TN with increasing rainfall (Figure 5), as reported in previous studies (e.g., Ide et al., 2007Ide et al., , 2012;;S. Inamdar et al., 2015;Lee et al., 2016;Taylor et al., 2015).However, exports of  NO3 − were higher than those of other nitrogen components in all rainfall events (Figure 5).The annual PN export, which included both fine particles (the constituents measured as PN in the present study) and coarse particles such as litter, was lower than the annual dissolved nitrogen export in 2001, a year in which there was a large-scale (417 mm) rainfall event (Osaka et al., 2011).These results imply that, in the Ab-S watershed, dissolved nitrogen is exported more readily than PN, even during extreme rainfall.
Previous studies of forest watersheds have shown that dissolved nitrogen exports exceed PN exports even during rainfall events (Ide et al., 2007(Ide et al., , 2012;;Lee et al., 2016), as in the present study, but others found the opposite, that is, that PN exports exceed dissolved nitrogen exports (S.Inamdar et al., 2015;Taylor et al., 2015;Wiegner et al., 2009).As shown in Figure 4, PN exports per unit streamflow increased with increasing streamflow in a rainfall event, whereas dissolved nitrogen exports increased almost linearly.Therefore, in a forest watershed where PN is readily exported, the increase in exported TN with increasing rainfall will be larger than in a forest watershed where dissolved nitrogen is readily exported.Conversely, in watersheds where the amount of dissolved  10.1029/2023WR034756 9 of 15 nitrogen exported exceeds the amount of PN exported during rainfall, as in this watershed, the increase in TN exported from the watershed in response to increases in rainfall will be small.
The low PN export in the Ab-S watershed in this study was due to low suspended solids (SS) export, rather than a low soil nitrogen concentration; the median nitrogen concentration at the soil depth of 0-10 cm was 0.36% (first quartile: 0.27%, third quartile: 0.44%) in the lower slopes and 0.31% (first quartile: 0.21%, third quartile: 0.40%) at the upper slopes, as shown in Figure 6.In S. Inamdar et al. (2015), in which PN export at the study site was high, the median nitrogen concentration was 0.40% in slope layer A and 0.45% in wetland layer A, while Taylor et al. (2015) reported a median concentration of 0.44% in the 0-15 cm soil surface layer.The soil nitrogen content was only slightly lower at Ab-S than at these other sites.However, the SS export amount at Ab-S (Table 1) was smaller than the amount reported by S. Inamdar et al. (2015), despite similar rainfall magnitudes, which may explain the low PN export in our study.
The slope of the streambed at Ab-S is ∼11.5° and that of the surrounding hills is 30°-40°; thus, in both cases the slope is steeper than at the study site of S. Inamdar et al. (2015).Consequently, the runoff rates (streamflow/ rainfall) at Ab-S were higher than those reported by S. Inamdar et al. (2015).However, the soil layer at Ab-S is thin, and there is little soil or sediment near the river channel because the bedrock (weathered granite) is exposed near the river.This would explain why SS and PN were not readily exported from our study area.Another reason was the relatively high nitrification rate in Ab-S soil.Previous studies showed that the net nitrification rate in soil is affected by the C:N ratio of the organic and soil layers (Aber et al., 2003;Emmett et al., 1998;Kuroiwa et al., 2011), and that soil mineralization and nitrification rates increase when the C:N ratio in the organic and soil layers is <20-25.The C:N ratio of the mineral soil in the Ab-S watershed was generally <20 at the upper and lower slopes (Figure 6d), and thus suitable for mineralization and nitrification.Therefore, the export of  NO3 − due to active mineralization and nitrification in the soil may explain the higher export of dissolved nitrogen than PN.Moreover, dissolved nitrogen export from Ab-S was higher than PN export, even during large rainfall events.Because the PN:dissolved nitrogen ratio in TN exported from the forest watershed strongly influenced the amount of TN export in response to increased rainfall, further comparative studies in other forest watersheds are needed.

Factors Controlling Nitrogen Export During Rainfall Events
Table 2 shows the coefficient of determination (R 2 ) and the slope of the regression lines between the logarithms of cumulative nitrogen export and those of rainfall, streamflow, peak flow, and direct runoff during rainfall events.PN had a higher R 2 with hourly peak flow, suggesting that particulate matter in the soil surface layer is readily released during short but intense rainfall events.The regression slopes between PN export and stream discharge or peak flow were >1, indicating that PN is more readily exported in large or intense rainfall events.However,  NO3 − and DON exports were close to 1 for streamflow and peak flow, which indicates minimal change in  NO3 − and DON exports per unit flow with rainfall magnitude and the absence of a sharp increase in response to intense rainfall events, in contrast to PN.However, regression analysis of these hydrologic parameters showed that the R 2 of  NO3 − exports was lower than that of PN and DON.This suggests that the amount of  NO3 − exported from the forest during rainfall is primarily influenced by hydrologic conditions but also by non-hydrologic factors, although the influence is smaller than for hydrologic conditions.
Figure 9 shows the relationship between streamflow and nitrogen concentrations in the stream water during rainfall.The concentration of DON in the stream water during rainfall was not closely related to the season or streamflow rate; it was relatively stable (approximately 0-0.5 mgN/L) compared with the PN and  NO3 − concentrations.The relationship between streamflow and PN concentration was also independent of season, although the PN concentration showed a clear tendency to increase at high streamflow rates.These results explain the strong correlation of DON export with streamflow, and of PN export with hourly peak flow.The relationship between  NO3 − concentration and streamflow differed by season: the  NO3 − concentration tended to be higher in summer and lower in spring and early summer, even with the same streamflow.This seasonal variation is consistent with the lower R 2 of  , 1994;Sebestyen et al., 2014).In Ab-S, the amount of litterfall was higher in the fall (Osaka et al., 2011), but there was no clear trend toward a low  NO3 − concentration in response to rainfall events.
Figure 10a shows the relationship between streamflow and  NO3 − export during rainfall events in summer and the rest of the year; Figure 10b shows the relationship between the amount of rainfall during the 7 days prior to a rainfall event and the ratio of measured  NO3 − export to  NO3 − export, calculated from the regression line in Figure 10a.Although  NO3 − export during summer events was within the prediction interval of the regression analysis of streamflow and  NO3 − export during events, it was frequently outside the confidence interval of the regression equation.This suggests that the amount of  NO3 − exported during summer rainfall events is greater than the average amount of  NO3 − exported during similar streamflow events. NO3 − export during rainfall events is also influenced by preceding rainfall (McDowell & Asbury, 1994).Specifically, when there is a large amount of rainfall prior to the  NO3 − export observation, some of the  NO3 − in the soil has already been washed out; the amount of  NO3 − export during the next rainfall event may be lower.However, in the summer rainfall in this study, the measured  NO3 − export was greater than the  NO3 − export calculated from the regression line, even after a large amount of preceding rainfall.Conversely, for many spring, fall, and winter rainfall events, the measured  NO3 − export was less than the  NO3 − export calculated from the regression line, regardless of the presence or absence of preceding rainfall.These results indicate the absence of a sharp drop in  NO3 − export in the Ab-S in summer, even if preceding rainfall partially washes out the  NO3 − in the soil.In contrast,  NO3 − export after rainfall events during spring, fall, and winter is strongly affected by  NO3 − washed out from the soil during preceding rainfall, such that the amount of  NO3 − exported during a subsequent rainfall is significantly reduced.As a result, for a rainfall event of the same magnitude in the Ab-S,  NO3 − export is more likely to occur during summer than the other seasons.The relationship between streamflow and the discharge-water-weighted mean concentration of SiO 2 at each rainfall event in this study is shown in Figure 11.In soil water and groundwater, SiO 2 is supplied by leaching from soil and rock minerals, such that groundwater with a longer residence time has a higher concentration of SiO 2 (Asano et al., 2003;Wels et al., 1991).Therefore, the SiO 2 concentration decreases during rainfall events when there is a large contribution of fast flow-paths, such as shallow subsurface and overland flows to stream water (Rose et al., 2018).Similarly, in the present study, the discharge-water-weighted mean concentration of SiO 2 decreased during rainfall events with large streamflows.This finding suggests that the contribution of fast flowpaths (e.g., shallow subsurface and overland flows) to stream water at the study site is also greater, especially during large rainfall events.However, there was no obvious seasonality in the relationship between streamflow and the discharge-water-weighted mean concentration of SiO 2 , implying that the contribution of fast flowpaths to stream water is mainly determined by the magnitude of rainfall and does not according to season.Therefore, seasonal differences in the relationship between streamflow and  NO3 − concentration shown in Figure 9 cannot be attributed to hydrological flowpaths, such as the contribution of shallow subsurface and overland flows to stream water.

Interaction Between Nitrogen Dynamics in Soil and 𝑨𝑨 NO𝟑𝟑 − Export in Rainfall Events
In general,  NO3 − exported from forest watersheds originated from either nitrification or the atmosphere.Several studies have shown that the  NO3 − exported from forest watersheds during baseflow and in response to rainfall events is nitrification-derived (Osaka et al., 2010;Spoelstra et al., 2001;Tsunogai et al., 2010).Therefore,  NO3 − exported from forests is likely influenced by  NO3 − production and consumption in soil, as well as hydrological processes.
At the present study site, the net nitrification rate in soil was strongly influenced by the soil temperature (Figure 7a); it was higher during the hot summer.In addition to soil temperature, the nitrification rate was affected by the soil water content and was higher at sites with a high weighted water content that coincided with a high soil temperature (Figure 8).As a result, the nitrification rate in the upper slopes, where nitrification does not usually occur because of the dry conditions, increased during the summer (i.e., when the soil water content was high) (Figure 8b).Nitrification rates in soils are reportedly significantly affected by also the C:N ratio of the forest floor or soil (Aber et al., 2003;Emmett et al., 1998;Kuroiwa et al., 2011), and by the soil water content (Avrahami & Bohannan, 2007;Ohte et al., 1997;Stark & Firestone, 1995).In the Ab-S, the C:N ratios of the soils at the lower and upper slopes are similar (Figure 6d), but the weighted water content of the soil is much lower at the upper slopes than at the lower slopes (Figure 6e); consequently, due to the dry conditions, the net nitrification rate in the upper slopes is usually low.However, heavy rainfall in the summer causes the upper slopes to become sufficiently wet enough to allow nitrification.Our study showed that, in summer, when the soil temperature is high and rainfall is abundant, nitrification can occur at both lower and upper slopes (due to the high humidity), resulting in an expanded area of nitrification, higher amount of  NO3 − produced at the watershed scale and higher amount of  NO3 − export per rainfall event than during spring, autumn or winter.This sequence of events is supported by the soil  NO3 − concentration in summer, which was as high or higher than during the other seasons despite the substantial washout of  NO3 − from the soil due to the abundant rainfall (Figure 7b).Matsumura et al. (2003) measured the seasonal variation of inorganic nitrogen uptake in a forest where the dominant tree species was Japanese cypress.The uptake of inorganic nitrogen was almost constant from spring to autumn; it tended to decrease in winter.Thus, in the Ab-S, the lack of intensive  NO3 − uptake by the cypress trees during summer may explain the high level of  NO3 − exported during a summer rainfall event.
The relationship between the seasonal variation in temperature and amount of rainfall and  NO3 − export during rainfall events at the study site is summarized in Figure 12.In summer (July-Sept.), rainfall is abundant and both soil temperature and water content in the lower and upper slopes are higher than during other seasons.As a result, nitrification in the soil is activated in the lower slopes, as well as the otherwise dry upper slopes, where nitrification is usually limited.Although plant uptake partly explains the uptake of  NO3 − from soil, the high nitrification activity during summer rainfall leads to  NO3 − exports that are higher than during other seasons (Figure 12c).Between fall and spring (October-March),  NO3 − uptake by plants is low; however, nitrifying activity in the soil is also low due to the small amount of rainfall and low soil temperature, such that only a small amount of  NO3 − is exported during rainfall (Figure 12a).During spring and summer (April-June), the soil temperature rises and nitrification activity starts to increase; because  NO3 − uptake by plants also increases, the amount of  NO3 − available for export during rainfall is low (Figure 12b).
One of the few studies that examined the linkage of plant-soil nitrogen cycling processes to  NO3 − export from forests during rainfall was conducted by Vaughan et al. (2017) in a Northeast American forest.The authors reported that  NO3 − export per streamflow was greater during the spring rainfall flow than during other seasons.The lower rate of  NO3 − export per streamflow in summer was attributed to more active  NO3 − uptake by plants; the lower rate in fall was attributed to the reduced nitrification, enhanced assimilation, and increased denitrification that occur in response to an increase in the dissolved organic carbon supply from fresh litter.
In contrast, the present study showed that  NO3 − export per streamflow is greater during the summer months in a mixed evergreen coniferous and broadleaf forest exposed to a monsoon climate, as mentioned above.The seasonal difference in the increase in  NO3 − export per streamflow during rainfall may be attributed to differences in climate and tree species between the two countries.Compared with the Northeast United States climate, the Japanese climate is wetter, especially during the warmer summer; this wetter climate promotes nitrification and increases the  NO3 − supply to the soil (Figures 7 and 8).Because the present study was conducted in an evergreen coniferous forest watershed, there is year-round litterfall (Osaka et al., 2011), although it is greater in the fall, and soil nitrogen uptake by vegetation also occurs throughout the year (Matsumura et al., 2003).Therefore, in contrast to the natural forests in the Northeast United States, where litterfall is concentrated in the fall and nitrogen uptake by vegetation is concentrated in the summer, the effects of fall litterfall and summer  NO3 − uptake by vegetation previously observed in the Japanese forest were small.It has previously been pointed out that  NO3 − is more readily exported during summer in Asian monsoon regions such as Japan, compared to "tight" Northeastern American forests where  NO3 − export is suppressed during summer when temperatures are higher and nitrification activity is more active (Chiwa et al., 2010;Mitchell et al., 1997;Ohte, 2012;Ohte et al., 2001;Ohte & Katsuyama, 2019).Ohte and Katsuyama (2019) analyzed the precipitation- NO3 − export relationship between growing (mainly in summer) and non-growing seasons of vegetation in Japanese forests based on the  NO3 − concentrations in stream water during baseflow and amount of precipitation of that period.There were no substantial seasonal differences in the precipitation- NO3 − export relationship in Japanese forests; the relationship remained linear.This indicates that hydrological conditions, mainly precipitation, are the driver of  (Creed & Band, 1998;Muraoka & Hirata, 1988;Ohte et al., 2003;Ohte & Katsuyama, 2019).Therefore,  NO3 − dynamics in surface soils are expected to be more strongly affected by rainfall, when the amount of water passing through the soil surface layer increases, than by baseflow.Therefore, through its focus on rainfall, this study was able to demonstrate a link between the hydrological and biogeochemical processes important for  NO3 − runoff from forests.These observations directly indicate that the increased  NO3 − export from forests in summer in Japan is the result of the high  NO3 − transport capacity from soil to stream water, as well as the high nitrification rate induced by the high temperature and humidity.Thus, in regions with high summer precipitation, such as the Asian monsoon region, the linkage between hydrological and biogeochemical processes leads to greater export of  NO3 − and therefore less N retention in the forest in summer.

Summary
Many studies have examined nitrogen export and fluctuations in the nitrogen concentration in stream water from forests in response to rainfall events.However, the effect of the seasonality of soil nitrogen dynamics on nitrogen export during small versus large rainfall events remains poorly understood.The findings of our study of the Ab-S watershed indicate that the coupling of biogeochemical processes, such as microbial activity in the soil, with hydrological processes regulates nitrogen export during rainfall.
Previous laboratory and in situ column experiments demonstrated that nitrification is enhanced under relatively high temperature and moisture conditions, whereas it is inhibited under dry conditions; the impact of temperature and moisture on  NO3 − leaching from soil has also been examined (Avrahami & Bohannan, 2007;Ohte et al., 1997;Stark & Firestone, 1995;Toda, 1994).Our study showed that the results of those studies can be linked to  NO3 − export from watersheds during rainfall events, which, although brief, are highly significant in terms of nutrient loss from forest ecosystems and nutrient supply to downstream ecosystems.Furthermore, comparisons with previous studies suggested that the effect of soil nitrogen dynamics on  NO3 − concentrations in stream water is stronger during rainfall than during base flow conditions.Thus, regions with wet summers due to rainfall, such as monsoon Asia, are more sensitive than other regions to  NO3 − export; the increased rainfall in summer has a particularly large impact on nitrogen export from forest watersheds.Our findings support a link between rainfall patterns of rainfall and temperature variability and the patterns of nitrogen retention in and nitrogen export from forests.
In the Ab-S watershed, when the amount of nitrogen export is expressed as an exponential function of streamflow during rainfall events, only PN had an exponent value that exceeded 1 (1.26); the values of the  NO3 − , DON, and TN exponents were approximately 1. Thus, even in extreme storms where rainfall exceeds 400 mm/day, nitrogen is exported without depletion, whereas PN is exported at an increased rate per stream volume, due to the heavy rainfall.However, in the present study, the total amount of PN did not exceed the total amount of  NO3 − , and the increase in TN export was approximately linear with respect to streamflow.In other forest watersheds, however, PN export exceeds dissolved nitrogen export (S.Inamdar et al., 2015;Taylor et al., 2015;Wiegner et al., 2009).Therefore, accurate predictions of nitrogen export from forests in response to climate change will require a detailed understanding of the factors that determine the ratio of particulate to dissolved nitrogen export from TN in watersheds differing in temperature and rainfall patterns during rainfall events.These factors may include a link between hydrological and biogeochemical processes for  NO3 − export from the watershed, as discussed above.We gratefully acknowledge the support of Dr. Etsuji Hamahata for setup of observation sites.We also acknowledge the support and suggestion of members of our laboratory at the University of Shiga Prefecture.This research was supported in part by a JSPS grant-in aid (nos. 25850114, 17K15288 and 20H04311) and Lake Biwa Environmental Research Institute (Research on Stream Water Quality Changes Associated with Forest Development, Forest Stream Water Monitoring).

Figure 1 .
Figure 1.(a) Locations of study site and (b) study area (Ab-S).

Figure 2 .
Figure 2. Mean monthly temperature and rainfall at the study site from 1998 to 2016.Error bars represent standard deviation.

Figure 6
Figure6shows the (a)  NO3 − concentration, (b) net nitrification rate, (c) soil nitrogen concentration, (d) soil C:N ratio, and (e) water content by weight (weighted water content) at soil depths of 0-10 cm and 20-30 cm at the upper and lower slopes of the watershed, respectively.The soil  NO3 − concentration and net nitrification rate were higher at the lower slopes than at the upper

Figure 4 .
Figure 4. Relationship between cumulative streamflow and cumulative nitrogen export during rainfall events.

Figure 6 .
Figure 6.Spatial distributions of (a)  NO3 − , (b) net nitrification rate, (c) nitrogen concentration, (d) C:N ratio, and (e) weighted water content in soil at study site Ab-S.

Figure 8 .
Figure 8. Relationship between the weighted water content and net nitrification rate at (a) lower slopes and (b) upper slopes.

Figure 7 .
Figure 7. Relationships between soil temperature and (a) the net nitrification rate and (b)  NO3 − concentration in soil.

Figure 10 .
Figure 10.(a) Relationship between streamflow and  NO3 − export during rainfall in summer and other seasons; (b) relationship between rainfall in the 7 days prior to the observation rainfall event and measured  NO3 − export relative to  NO3 − export calculated from the regression line in (a).Blue dotted line indicates confidence interval and orange dotted line indicates prediction interval (calculated based on 90%).

Figure 11 .
Figure 11.Relationship between streamflow and the discharged-waterweighted mean concentration of SiO 2 during each rainfall event.

Figure 12 .
Figure 12.Conceptual model of the relationship between  NO3 − dynamics in soil and  NO3 − export from a forest watershed during rainfall events.

Table 1 Results
of Rainfall Event Survey Figure 3. Relationship between rainfall and direct runoff during rainfall events.

Table 2
Relationships (R 2 ) Between Logarithm of Cumulative Nitrogen Export During Rainfall Events and Logarithm of Hydrological Parameters NO3 − than PN or DON with streamflow and rainfall amount.In North and Central American forests, the  NO3 − concentrations in stream water decreased during the litterfall season, especially under baseflow conditions (McDowell & Asbury NO3 − export from forests, and that  NO3 − export from forests does not depend on the amount of  NO3 − in the soil but is determined almost entirely by the  NO3 − transportation capacity from the soil.However, the present study focused on observations during rainfall, which is vital for  NO3 − export from forests, and found that not only does  NO3 − export increase with soil  NO3 − transport capacity (magnitude of rainfall) during rainfall, but that  NO3 − export per streamflow is higher during summer rainfall than during other seasons.The fact that we were able to find seasonal differences in  NO3 − export per streamflow by using detailed data obtained through observations during rainfall, as in the latter case, suggests that  NO3 − dynamics in soil influence  NO3 − concentrations in stream water during rainfall, rather than during baseflow.Notably, the findings are consistent with the conceptual model of  NO3 − export from forested watersheds established in previous studies, whereby seasonal changes in  NO3 − runoff concentrations and increases in  NO3 − concentrations during rainfall could be explained by the increased contribution of relatively  NO3 − rich water passing through the soil surface layer during higher flow events