3-1. Pattern of nitrate leakage
Figure 3-1 shows the change in nitrate concentration in the Hyakumakidani stream over approximately 24 years, from July 1998 to December 2022. Nitrate concentrations fluctuated greatly according to the stream’s water flow (Kawakami et al., 2001).
Figure 3-2(a) shows the nitrate concentration change by flow rate in the Hyakumakidani stream from August 1998 to July 1999. The nitrate concentration increased with the increase in flow rate. The phenomenon, in which the nitrate concentration increases with an increase in the flow rate, has been reported in many rivers (Baurès et al., 2013). Sase et al., Kamisako et al., Mitchell et al., and Shibata et al. reported that the nitrate concentration increased with an increase in the flow rate at a site in Japan where nitrogen saturation is suspected (Sase et al., 2019, Kamisako et al., 2008, Mitchell et al., 1997, Shibata et al., 2001). However, the unique characteristic of the Hyakumakidani stream is that a linear relationship was obtained when the logarithm of the flow rate was taken, as shown in Figure 3-2(b). Linearity was maintained with a 100-fold increase in the flow rate from approximately 0.1 L/s to 10 L/s. This indicates that the nitrate concentration seemed to be determined by the flow rate and was not affected by the season. To clarify the independence of the nitrate concentration from the season, nitrate concentrations in the growing season and the dormant season were considered separately.
Figure 3-3 (a) shows the tree growth as measured by dendrometers attached to 6 deciduous stands and 6 coniferous stands randomly selected on the Hyakumakidani watershed. The average relative trunk diameters of both the coniferous and deciduous stands against those of April 2000 are indicated in the figure. Since both coniferous and deciduous stands grew between May and August, the plots in Figure 3-2 (b) were separately plotted again in Figure 3-3 (b) for the growing season (May to August) and for the dormant season (September to April). There was no difference in nitrate concentration between the growing season and the dormant season. This meant that the Hyakumakidani stream was in Stage 3 of nitrogen saturation, according Stoddard’s definition, in which there are no seasonal changes in nitrate concentration, and the nitrate concentration exceeds that of precipitation (Stoddard, 1994).
Mitchell et al. reported that Japanese streams containing high concentrations of nitrate generally had less fluctuation in nitrate concentration. They pointed out the Japanese meteorological features, in which, during the growing season, substantial precipitation combined with high temperature enhances nitrogen mineralization and nitrate leakage to stream water, causing less nitrate fluctuation (Mitchell et al., 1997). On the contrary, Kureha Hill is located on the Japan Sea side of Japan, which receives a large amount of snowfall in the winter. Different from the watersheds on the Pacific Ocean side of Japan investigated by Mitchell et al., the flow rate in the dormant season is generally higher than that in the growing season. Therefore, there should be other reasons for the low fluctuation in nitrate concentration. One reason may be less utilization of nitrogen than in the nitrogen cycle of the forest ecosystem of Kureha Hill. Makino et al. conducted a statistical analysis of the nitrate concentration of stream water in the Kinki District to compare the nitrate concentrations in the regions on the Japan Sea side and on the Pacific Ocean side. They concluded that the lower nitrate concentration on Japan Sea side than that on Pacific Ocean side is caused by the flushing of nitrogen in the soil and dilution by heavy rain and snow on Japan Sea side (Makino et al., 2021, Makino et al., 2023). Since Kureha Hill is located on the Japan Sea side and receives a large amount of snow, the flushing of nitrogen and dilution could take place. However, the nitrate concentrations in the streams are quite high.
Figure 3-4 shows the relationship between the flow rate and nitrate concentration in the Hyakumakidani stream in 2019. The trend for nitrate concentrations to increase as the flow rate increased was the same; however, a linear relationship was observed between the nitrate concentration and flow rate without a logarithm in 2019.
Figure 3-5 shows the change in the pattern of nitrate leakage according to the flow rate. When the flow rate was less than 7 L/sec., the nitrate concentration in 1999 exceeded that in 2019. The nitrate concentrations in 1999 and 2019, at an annual average flow rate of 1.4 L/sec in 2009, were 106 µeq/L and 69.4 µeq/L, respectively. This indicates that the nitrate concentration decreased to 65.5% in 2019, as compared with the average flow rate in 2009.
Since the nitrate concentrations were a function of the flow rate, and seasonal variations were not found in either 1999 or 2019, the nitrate concentrations in 1999 and 2019 were estimated from the flow rate, and the nitrate leaching was calculated.
The nitrate concentration in each year was estimated by using the relationship between the flow rate and the nitrate concentration in 1999 and 2019, as shown in Figures 3-2 and 3-4, respectively. In order to compare the nitrate concentrations with the same flow rate pattern, the flow rate in 2009 was used as an example, since the precipitation in 2009 was 2224 mm which was equivalent (93%) to the average precipitation (Japanese Meteorological Agency). Figure 3-6 shows the flow rate in 2009 and the estimated nitrate concentrations by 1999 and 2019. According to the equation, the relationships were estimated to be 2175 mol/ha/year and 1887 mol/ha/year in 1998 and 2019, respectively. The nitrate leakage was reduced to 87% over the 20-year period, which was not as low as the reduction rate of 65.5% when it was compared at the average flow rate. This is because the nitrate ion concentration given by the 2019 model formula at high flow rates (>7 L/sec) exceeded the 1999 model formula, and the 2019 model’s high flow rate and high nitrate concentration gave a large quantity of nitrate lodgings. In 2009, a high flow rate of more than 7 L/sec. took place several times in July and November, which caused high loading of nitrate.
Figure 3-7 compares the nitrate concentrations at each site on Kureha Hill between 2005–2006 and 2019. Flow rates at the Hyakumakidani stream were used for comparison, since flow rates were measured only at the Hyakumakidani stream. For 2019, the nitrate concentrations at each site are shown when the flow rates in the Hyakumakidani stream were 0.12, 1.65, and 6.90 L/sec. For 2005–2006, nitrate concentrations at these flow rates were estimated from the equations of the relationship between the flow rate and nitrate concentration in 2005–2006. As mentioned previously, in 2009, the average annual flow rate in the Hyakumakidani stream was 1.4 L/sec. Therefore, the flow rate of 0.12 L/sec. corresponds to a low flow rate, 1.65 L/sec. corresponds to the average flow rate, and 6.90 L/sec. corresponds to a high flow rate. As a whole, nitrate runoff concentrations in 2019 were less than those in 2005–2006, when compared at the same flow rate. It is apparent that the nitrogen input to the catchment by deposition or addition is one of the factors that regulates the nitrate concentration in the stream water (Tietema et al., 1997, Nishina et al., 2017, Ohrui & Mitchell, 1997, Dise & Wright, 1995). Some reports demonstrated a decrease in the nitrate concentration of stream water with a decrease in atmospheric nitrogen deposition due to recent air pollution control measurements (Eshleman et al., 2013, Chiwa, 2021, Gilliam et al., 2019, Baba et al., 2020).