Monitoring of Non-point Source Pollutants Generated by a Flower Farm

Abstract

This paper considers the effect of rainfall on non-point source (NPS) pollutant loads. The impact of runoff on the occurrence of NPS pollutants was found to be influenced by rainfall amount, rainfall intensity, and the number of antecedent dry days (ADD), both independently and in combination. The close correlation ($R^2$ = 0.9920) between rainfall and runoff amounts was demonstrated at the study site (a flower farm) over the period between January 2011 and December 2013. The relationships among pollutant levels, runoff, and rainfall was not satisfactory results except for the Biochemical Oxygen Demand ($BOD_5$). The correlation coefficients between $BOD_5$, and both runoff and rainfall, were greater than 0.92. However, the relationships of other pollutants, such as Suspended Solid (SS), Chemical Oxygen Demand ($COD_{Mn}$), Total Nitrogen (TN), and Total Phosphorus (TP), with runoff and rainfall had correlation coefficients of less than 0.70. The roles of rainfall was different from rainfall categories on the occurrence of runoff. Instantaneous rainfall intensity was a principle factor on the occurrence of runoff following light rainfall events (total ${\leq}30mm$). For rainfall of intermediate intensity (total precipitation 31-50 mm), the combined effect of both average rainfall intensity and ADD was found to influence runoff generation. We conclude that the control of NPS pollutants with the reflection of the climate change that makes the remarkable effect of amounts and forms on the rainfall and runoff.

Introduction

Efforts to improve water quality in the watersheds have progressed well with respect to point source pollutants, but the management of NPS pollution remains a significant challenge. Water quality and pollutant loads in rainfallderived runoff from cropland is not only influenced by climatic factors such as precipitation, rainfall intensity, and the number of preceding dry days, but also by agricultural factors such as the characteristics of the crops being cultivated (Oh et al., 2004; Shin et al., 2006; Kang et al., 2009; 2010a; 2010b; Ryu et al., 2011; Yoon et al., 2011; Kim et al., 2013; Choi and Kang., 2014). In addition, as a result of the recent increase in the frequency of droughts and flooding, it is be coming much more difficult to manage NPS pollutants in agricultural areas. It has been reported previously that the three major types of NPS pollutants in agricultural areas are sedimentation, nutrient salts, and pesticide runoff (Carpenter et al., 1998; Hunt et al., 1999; Schultz, 2004). The problems associated with increases in NPS pollutants from agricultural areas are also increasing overseas. In developing countries, the use of water is becoming more important because of thriving agricultural activities and health considerations, whereas in developed countries (Reddy and Behera, 2006), with the successful reduction of pollutants from point sources, NPS pollutants from agricultural areas are reported to contribute the highest proportion to total water pollution (Environmental Protection Agency, 2000). Rainfall-derived runoff from agricultural areas can contaminate surface water in various ways, and can be categorized into two types. One is direct runoff, through agricultural drainage into surface waters such as rivers, and the other is seeping underground, which creates cracks as source of water in the ground, or may directly affect the surface water contamination (Brian et al., 2008). Horan et al. (2002) attempted to develop an approach to managing both point source pollutants and NPS pollutants more economically, and proposed the combination ratio of the two pollutants as the most influential factor in efforts to meet planned water quality standards. In addition, they concluded that, because NPS only affect water quality in small watersheds, exchange programs between two (point and non-point)would be in effective. Although NPS pollutants are difficult to control because they are generated by runoff from rainfall that shows extreme temporal and spatial variations, it is widely accepted that they are the most significant factor in terms of achieving water quality improvements. This study aims to determine the impact of runoff from NPS on water pollution loads by identifying the major factors that affect the generation of rainfall-derived runoff. Rainfall data were obtained from monitoring conducted over the past three years at a floriculture site. We believe that our findings will contribute to the effective management of rainfallderived runoff from NPS.

 

Methods

The flower farm used here as the monitoring site for NPS pollutants is typical agricultural land (Andong, South Korea: N36°32'37.8'', E128°47'17.8''), and covers an area of 631.3 m2 with a slope of 8.5%. A flower farm was selected as one of NPS in the classification scheme of the Ministry of Environment, Korea (2014a). Rainfall, runoff, and the pollutant load in the runoff were monitored at the farm between January 2011 and December 2013. Rainfall amounts and runoff rates were measured on a per-minute basis using a raingauge (Environdata weather station rain gauge, RG 20, Australia) and a flow meter (Flo-Tote3, USA), respectively, from the start of each rainfall event. However, for a very small amount of rainfall runoff where the measurement on per-minute basis was not possible, spills were taken into a beaker and runoff speed was calculated by dividing by the corresponding time. Runoff samples were collected by the National Institute for Environmental Research (NIER, 2012). The runoff samples were analyzed for SS, BOD5, CODMn, TN and TP according to the procedures of the Ministry of Environment (2014b). The EMC (event mean concentration) and TPM (total pollutant mass) were calculated using the equations 1 and 2 (NIER, 2010, 2012).

 

Results and discussion

Characteristics of precipitation

Over the monitoring period, rainfall fell a total of 310 times at the survey site, giving a total accumulation of 3279 mm. Rainfall data are shown in Table 1. Rainfall events of less than 10 mm accounted for 69.4% of all events, but only contributed 16.1% to the total rainfall accumulation, but it is difficult to identify any significant impact on the surface water from the classified data (Table 1). Thus, This suggests that the pollutants remain fixed within the farmland soil. As for the monitoring rate by rainfall class, 100% was performed in the range of ‘less than 10 mm’, 18% in ‘10 to 30 mm’, 28% in ‘30 to 50 mm’, 45% in ‘more than 50 mm’.

Table 1.*a : Rainfall amounts for each rainfall category/total rainfall amounts. b : Number of rainfall events monitored for each rainfall category/total number of rainfall events for each rainfall category.

Rainfall characteristics influencing occurrence of runoffs

The characteristics of the monitored rainfall events are shown in Table 2 and Fig. 1. As rainfall in the < 10 mm category did not generate any runoff during the monitoring period, only the results for rainfall exceeding a total of 10 mm are shown in Table 2 and Fig. 1. In terms of monitored rainfall, there were 22 rainfall events exceeding 10 mm, with amounts of rainfall between 11.8 and 156.4 mm, and the number of antecedent dry days (ADD) was in the range 0.9 to 32 days. However, rainfall events in the ≤ 10 mm category were included in the count of dry days when calculating ADD.

Table 2.*Zero = no runoff.

Fig. 1.Precipitation factors affecting runoff.

For Event 20, 11.8 mm of rainfall, 0.9 days (22 hours) of ADD, and 7.9 mm of average rainfall intensity were recorded, and this was the shortest ADD and highest average rainfall intensity of all monitored rainfall events; however, no runoff was recorded. This suggests that rainfall was a more important control on the occurrence of runoff than average rainfall intensity or ADD. The correlation coefficient (R2) between rainfall and runoff came out with avery significant value of 0.9920 for 10 of the rainfall events where runoff occurred, and it can be seen that runoff also increases as rainfall increases (Fig. 2). On the other hand, the R2 value between rainfall and runoff calculated from previous research at a potato farm was 0.6734 (Kang et al., 2009). The range of rainfall was less than 100 mm at this potato farm, and the correlation between rainfall and runoff was lower than that found here.

Fig. 2.Positive correlation between rainfall and runoff amounts.

According to previous research by Kang et al. (2009, 2010b) on similar cropland close to the present study site, the average rainfall intensity is an important factor in the development of runoff for rainfall events in the ≤ 20 mm category. We recorded nine events in this ≤ 20 mm category (Fig. 3), but only two that generated runoff, namely Event 13 (rainfall total = 18.0 mm) and Event 17 (rainfall total = 14.6 mm). For Event 13, the average rainfall intensity was 3.2 mm/hr, but the instantaneous rainfall intensity was the highest in this category (≤ 20 mm) at 47.3 mm/hr. We interpret this to show that runoff occurred because of the effect of the momentarily increased rainfall intensity. For Event 17, we also believe that the small amount of runoff (0.05 m3) was generated by the relatively high instantaneous rainfall intensity (16.0 mm/hr), although the average rainfall intensity during this event was lower at 0.8 mm/hr. For Events 06 (rainfall total = 11.8 mm) and 20 (rainfall total = 11.8 mm), we believe that runoff did not occur because these events experienced the lowest rainfall totals in the ≤ 20 mm category, although the instantaneous rainfall intensities were 18.8 and 23.4 mm/hr, respectively. These observations suggest that instantaneous rainfall intensity is the most important control on the development of runoff for events in the ≤ 20 mm category.

Fig. 3.Precipitation and runoff in the ≤ 20 mm category. *Average R.I. : Average rainfall intensity during the event. Instantaneous R.I. : Instantaneous maximum rainfall intensity during the event.

For Events 02 and 22, runoff might have been expected to occur because rainfall totals were 37.6 and 41.4 mm, respectively; however, no runoff was recorded. For Event 02, the ADD was 32 days, the longest among all of the rainfall events monitored here, and this long ADD probably prevented runoff from occurring. For Event 22, we suggest that the lack of runoff was caused by the influence of the ADD (12 days) and average rainfall intensity (1.2 mm/hr). On the other hand, for Event 11 (rainfall total = 45.2 mm, ADD = 17 days) and Event 14 (rainfall total = 38.4 mm, ADD = 21 days), which were similar to Event 22, runoff was recorded. This is because the average rainfall intensity in Events 11 and 14 was higher than that of Event 22, at 1.8 and 2.2 mm/hr, respectively. Therefore, it can be seen that what influenced the occurrence of runoff in the 30-50 mm class was not a single factor, but the combined effect of multiple factors including average rainfall intensity and ADD (Fig. 4).

Fig. 4.Factors affecting runoff in each rainfall category.

Events 15 (rainfall total = 81.8 mm, ADD = 1 day) and 19 (rainfall total = 78.0 mm, ADD = 1 day) showed a large differences in runoff with rates of 0.101 (runoff = 5.2 m3) and 0.026 (runoff = 1.3 m3), respectively, being recorded, although both events generated a similar amount of rainfall and had the same ADD value. Event 15 had higher average and instantaneous rainfall intensities than Event 19, with an average rainfall intensity of 5.2 mm/hr and an instantaneous rainfall intensity of 64.9 mm/hr, whereas Event 19 had an average rainfall intensity of 3.0 mm/hr and an instantaneous rainfall intensity of 9.5 mm/hr. Therefore, we suggest that the rate of runoff was higher in Event 15 because of the influence of the relatively higher average and instantaneous rainfall intensities. These results show that the amount of runoff was strongly affected by the average and instantaneous rainfall intensity. In addition, as runoff varied among events within the same rainfall class, depending on the correlation (R2) between two parameters (runoff and rainfall), it can be seen that current NPS management methods used to estimate rainfall pollutant loads by classifying the rainfall events in terms of rainfall amounts have some limitations.

Pollutant runoff characteristics

The EMC values of the pollutants by rainfall event are shown in Table 3. The EMC of each pollutant was as follows: SS (6.31 × 10 mg/L ~ 1.27 × 104mg/L), BOD5 (2.87 mg/L ~ 9.98 mg/L), CODMn (1.67 × 10 mg/L ~ 1.50 × 102 mg/L), TN (8.38 × 10−1 mg/L ~ 1.97 × 10 mg/L), TP (7.87 × 10−1 mg/L ~ 1.06 × 10 mg/L). The pollutant with the largest range of EMC was SS, and the least variations in the concentration were shown in BOD5.

Table 3.*t N = × 10

As shown in Table 4, the pollutant loads by rainfall event were as follow: SS (1.21 × 10 g ~ 1.40 × 105 g), BOD5 (1.11 × 10−1g ~ 1.43 × 102 g), CODMn (5.23 × 10−1 g ~ 2.58 × 103 g), TN (1.31 × 10−1 g ~ 2.97 × 102 g), TP (7.22 × 10−2g ~ 1.47 × 102 g). As for variation in the pollutant loads, SS, CODMn and TP were at the level of 104 while BOD5 and TN were at the level of 103.

Table 4.*t N = × 10

The correlation of the pollutant loads with rainfall and runoff are shown in Fig. 4. Rainfall and BOD5 show the highest correlation (R = 0.92), followed by rainfall-TN (0.71), rainfall-CODMn (0.70), rainfall-TP (0.67), and rainfall-SS (0.58). For the relationship between runoff and pollutant loads, runoff-BOD5 showed the highest correlation of 0.99, followed by runoff-TN (0.68), runoff-CODMn (0.66), runoff-TP (0.63), and runoff-SS (0.53).

Fig. 5.Relationship among pollutant load, runoff and rainfall.

As correlation levels between rainfall and runoff and pollutant loads did not vary among the individual pollutants, a higher correlations were not observed in each pollutant except in BOD5. In contrast, correlations between the pollution loads of SS and other pollution loads were shown higher with the SS-BOD5 (R = 0.57), SS-CODMn (0.99), SS-TN (0.98), SS-TP (0.99). Therefore, when developing predictive models for NPS management of the agricultural area in the future, it is considered to be effective to establish a baseline by using the pollution loads of SS.

 

Conclusions

During the monitoring period, the highest rainfall total recorded was 156.4 mm, and this generated runoff of 38.80 m3. The relationship between runoff (y) and the total amount of rainfall in an event (x) can be expressed by the equation y = 0.0029x2 − 0.2350x + 4.1342 (R2 = 0.9920). However, class-specific factors (i.e., related to the total amount of rainfall)were found to influence runoff, either singly or in combination. Looking at the factors affecting runoff by rainfall class, instantaneous rainfall intensity only played a role as a single factor in the lower rainfall range (≤ 30 mm), whereas the average rainfall intensity and ADD acted together at intermediate rainfall amounts (31-50 mm). In the higher range (> 50 mm), runoff occurred for all rainfall events. On the other hand, as a result of the influence of rainfall intensity, the rate of runoff varied considerably between similar levels of rainfall. Therefore, the current method of calculating NPS pollutant loads on the basis of a single factor classified by rainfall class has proven to be very limited in its effectiveness.

Variations in the concentration of pollutants in runoff are experiencing dynamic changing process, with SS on the level of 103 and others on the level of 102 while correlation is shown higher with more than 0.98 in loads of SS and those of BOD5, CODMn, TN and TP (except 0.57 in the BOD5). Therefore, it is expected that the pollution level in rainfall-generated runoff can be reduced through the management of soil effluents causing the increase in SS.