**4. Results and discussion**

#### **4.1. Results of base flow separation**

After collecting the annual stream flow data of the stations, the PART program developed by the United States Geological Survey (USGS) was used for analysis and to calculate the base flow of each station. The results showed that the trends of the base flow and stream flow were consistent, with the former ranging from 60.9 to 284.9 cm/year. Since Taiwan is located in a subtropical humid zone with perennial precipitation, during the steady base flow period, the recharge derived from base flow separation might still be greater than the groundwater recharge and discharge situations reflected by stream flow.

From a short‐term perspective, base flow is affected by the amount of precipitation for that year. However, the amount of water discharged from the groundwater system should be a constant value over the long‐term. Hence, stable base flow analysis was used to examine and estimate the groundwater recharge. Lee et al. [7] pointed out that the calculated base flow is likely to be overestimated when the base flow estimation method is used to assess stream flow data. This will in turn affect the calculations of the groundwater recharge. After estimating the base flow for all eight stations, evaluation was made using stable base flow analysis. This provided the depth of annual groundwater recharge for each station. The results are shown in **Table 2**. The depth of annual groundwater recharge for the northern region was 15.6–160.65 cm/year. The maximum and minimum values were recorded at the Fushan and Niudou stations, respectively.


**Table 2.** Results of the depth of annual groundwater recharge in Northern Taiwan.

#### **4.2. Analysis of groundwater recharge**

**Figure 2.** Spatial distribution of gauging stations in Northern Taiwan.

recharge and discharge situations reflected by stream flow.

After collecting the annual stream flow data of the stations, the PART program developed by the United States Geological Survey (USGS) was used for analysis and to calculate the base flow of each station. The results showed that the trends of the base flow and stream flow were consistent, with the former ranging from 60.9 to 284.9 cm/year. Since Taiwan is located in a subtropical humid zone with perennial precipitation, during the steady base flow period, the recharge derived from base flow separation might still be greater than the groundwater

From a short‐term perspective, base flow is affected by the amount of precipitation for that year. However, the amount of water discharged from the groundwater system should be a constant value over the long‐term. Hence, stable base flow analysis was used to examine and estimate the groundwater recharge. Lee et al. [7] pointed out that the calculated base flow is likely to be overestimated when the base flow estimation method is used to assess stream flow data. This will in turn affect the calculations of the groundwater recharge. After estimating the base flow for all eight stations, evaluation was made using stable base flow analysis. This provided the depth of annual groundwater recharge for each station. The results are shown

**4. Results and discussion**

**4.1. Results of base flow separation**

34 Groundwater - Contaminant and Resource Management

After calculating the depth of groundwater recharge for the respective stations using stable base flow analysis, the results were multiplied by the area of the water catchment. This gave the annual groundwater recharge for each station and the long‐term average amount (**Table 3**). Annual groundwater recharge for the northern region worked out to be between 2.67 × 10<sup>7</sup> and 7.81 × 10<sup>8</sup> m3 /year. Both largest and smallest volumes were found in the Tamsui river basin, at the Gaoyi and Hengxi stations, respectively.


**Table 3.** Results of annual groundwater recharge in Northern Taiwan.

#### **4.3. Analysis of trends in groundwater recharge**

The Mann‐Kendall test was used to analyze the characteristics of the long‐term annual recharge trends for Northern Taiwan and to examine the distribution of significant trends. The test results are shown in **Table 4**. The significant level α = 5% was adopted as the standard, meaning that |*Z*α/2| = 1.96 was used to verify that a trend was of significance. Three river flow stations were found to exhibit significant trends, namely Fushan, Hengxi, and Ximen Bridge stations. It was a significant upward trend for the first two stations, but a significant downward trend for the third station (**Table 4**). Overall, among the eight stations within the study area, only two exhibited downward trends: Gaoyi station at upstream Tamsui River and Ximen Bridge at downstream Lanyang River. The spatial distribution of trends for the various stations is shown in **Figure 3**. In addition, the Theil‐Sen estimation method was used to calculate the trend slope. An upward and downward trend is indicated by a value larger and smaller than zero, respectively. The trend line can also be calculated using the trend slope values and annual recharge volume for the individual years. **Equation (10)** was then used to calculate the rate of change and the results are shown in **Figure 4**.


Note: \* indicates the significant trends. The positive values represent increasing trends, and the negative ones represent decreasing trends.

**Table 4.** Results of significant trend, slope, and relative change in Northern Taiwan.

It was observed that the rate of change for the annual recharge was small at the Fengshan and Touqian river basins: that for the Xinpu, Neiwan, and Shangping stations was all less than 10%. For the Tamsui river basin, with the exception of Gaoyi, the rate of change for the other stations was greater than 30%. In particular, the increases for Fushan and Hengxi stations were 44.0 and 33.7%, respectively. The rate of change was also greater than 30% for the Niudou and Ximen Bridge stations. However, it was an increase of 30.9% for the former, but a decrease of 106.6% for the latter. The slopes of the significant trends in stream flow are shown in **Figure 4**. Spatiotemporal Analysis of Groundwater Recharge Trends and Variability in Northern Taiwan http://dx.doi.org/10.5772/63508 37

**Figure 3.** Map showing spatial variation in trends in annual recharge in Northern Taiwan.

**4.3. Analysis of trends in groundwater recharge**

36 Groundwater - Contaminant and Resource Management

change and the results are shown in **Figure 4**.

Note: \*

decreasing trends.

**Basin Station Record years Mann-Kendall**

The Mann‐Kendall test was used to analyze the characteristics of the long‐term annual recharge trends for Northern Taiwan and to examine the distribution of significant trends. The test results are shown in **Table 4**. The significant level α = 5% was adopted as the standard, meaning that |*Z*α/2| = 1.96 was used to verify that a trend was of significance. Three river flow stations were found to exhibit significant trends, namely Fushan, Hengxi, and Ximen Bridge stations. It was a significant upward trend for the first two stations, but a significant downward trend for the third station (**Table 4**). Overall, among the eight stations within the study area, only two exhibited downward trends: Gaoyi station at upstream Tamsui River and Ximen Bridge at downstream Lanyang River. The spatial distribution of trends for the various stations is shown in **Figure 3**. In addition, the Theil‐Sen estimation method was used to calculate the trend slope. An upward and downward trend is indicated by a value larger and smaller than zero, respectively. The trend line can also be calculated using the trend slope values and annual recharge volume for the individual years. **Equation (10)** was then used to calculate the rate of

**test result**

Ximen Bridge 1983-2012 -3.604 -5.254 -106.6%

Gaoyi 1957-2002 -0.095 -0.026 -0.8% Hengchi 1958-2012 2.195 0.363 33.7%

Shang-Ping 1971-2012 0.356 0.105 7.9%

indicates the significant trends. The positive values represent increasing trends, and the negative ones represent

It was observed that the rate of change for the annual recharge was small at the Fengshan and Touqian river basins: that for the Xinpu, Neiwan, and Shangping stations was all less than 10%. For the Tamsui river basin, with the exception of Gaoyi, the rate of change for the other stations was greater than 30%. In particular, the increases for Fushan and Hengxi stations were 44.0 and 33.7%, respectively. The rate of change was also greater than 30% for the Niudou and Ximen Bridge stations. However, it was an increase of 30.9% for the former, but a decrease of 106.6% for the latter. The slopes of the significant trends in stream flow are shown in **Figure 4**.

Lanyang River Niu-Dou 1979-2010 1.768 0.155 30.9%

Danshui River Fu-Shan 1953-2012 2.991 1.198 44.0%

Fengshan River Hsin-Pu 1970-2012 0.415 0.049 10.0% Touqian River Nei-Wan 1971-2012 0.398 0.087 7.0%

**Table 4.** Results of significant trend, slope, and relative change in Northern Taiwan.

**Slope estimator Relative change**

The change‐point analysis results revealed that change points occurred at three gauging stations: Fushan, Hengxi, and Ximen Bridge (**Table 5** and **Figure 5**). The change point for the Fushan station occurred in 1999. The average annual groundwater recharge before and after the change point was 23.1 × 10<sup>2</sup> and 35.2 × 10<sup>7</sup> m3 /year, respectively. The amount exhibited an upward trend post‐1999, and the rate of increase was 46.9%. For the Hengxi station, the change point occurred in 1983, with the before and after volume being 2.2 × 107 and 3.3 × 10<sup>7</sup> m3 year, respectively. There was an overall upward trend as well, and the rate of increase was 41.7%. The change point for the Ximen Bridge station occurred in 2001. The before and after volumes there were 19.6 × 10<sup>7</sup> and 6.7 × 10<sup>7</sup> m3 /year, respectively. The overall trend after the change point declined by as much as 89.4%. Among the three aforementioned gauging stations, the magnitude of change at the Ximen Bridge station was the largest.

**Figure 4.** The significant trend line of gauging stations in Northern Taiwan. (a). Fu‐Shan station, (b) Hengchi station, (c) Ximen Bridge station


Note: The number in bold indicates a statistically significant difference.

**Table 5.** Results of change points using cumulative deviations and Mann-Whitney-Pettitt test.

Spatiotemporal Analysis of Groundwater Recharge Trends and Variability in Northern Taiwan http://dx.doi.org/10.5772/63508 39

**Figure 5.** The results of change points of significant trend stations. (a)Fu‐Shan station, (b) Hengchi station, (c)Ximen Bridge station

#### **5. Conclusions**

**Figure 4.** The significant trend line of gauging stations in Northern Taiwan. (a). Fu‐Shan station, (b) Hengchi station,

**Mean recharge (107 m3**

**after change point**

2001 0.9999 19.6 6.7 -89.4%

**/year)**

**Relative change at the change point**

**P (Mann-Whitney-Pettitt)**

**before change point**

Lanyang River Niu-Dou - 0.8423 - - -

Danshui River Fu-Shan 1999 0.9995 23.1 35.2 46.9%

Fengshan River Hsin-Pu - 1.0660 - - - Touqian River Nei-Wan - 0.8012 - - -

**Table 5.** Results of change points using cumulative deviations and Mann-Whitney-Pettitt test.

Note: The number in bold indicates a statistically significant difference.

Gaoyi - 0.8553 - - - Hengchi 1983 0.9990 2.2 3.3 41.7%

Shang-Ping - 0.7013 - - -

(c) Ximen Bridge station

**Basin Station Change**

38 Groundwater - Contaminant and Resource Management

Ximen Bridge **point (year)**

> In this study, the base flow estimation method and stable base flow analysis were used to evaluate the long‐term stream flow data of eight stations in Northern Taiwan. The range was between 2.67 × 107 and 7.81 × 108 m<sup>3</sup> /year based on the cumulative annual recharge. The characteristics of the trends were examined using the Mann‐Kendall test. Further, trend slope calculation and change‐point analysis were carried out. The results provided an understanding

of the trends for the eight stations and indicated that only two had downward trends: Gaoyi station at the upstream of the Tamsui River and Ximen Bridge at the downstream of the Lanyang River. The rate of change for the Fengshan and Touqian river basins was relatively small. Separately, the rate of change for the annual groundwater recharge was greater than 30% for all stations in the Lanyang river basin, while Gaoyi station was the only exception in the Tamsui river basin. The results of the change‐point analysis showed that the change point for the Fushan, Hengxi, and Ximen Bridge stations occurred in 1999, 1983, and 2001, respec‐ tively. The average annual groundwater recharge for the first two stations exhibited an upward trend before and after the change point (46.9 and 41.7%, respectively), while that for the last station decreased by as much as 89.4%. The findings can be used for regional hydrological studies and as reference for water resources planning.
