Academia.eduAcademia.edu

Figure 7 – uploaded by Rene Lefebvre

See full PDFdownloadDownload figure
Fig. 6 Mean monthly values obtained with hydrograph separation (J) and the HELP model (W, before correction) representative baseflow increases when total flow is high. Therefore, annual recharge estimates obtained using 12 months seem largely overestimated in this case. Similar conclusions were drawn by Croteau et al. (2010) for an area also located in eastern Canada.

Figure 6 Mean monthly values obtained with hydrograph separation (J) and the HELP model (W, before correction) representative baseflow increases when total flow is high. Therefore, annual recharge estimates obtained using 12 months seem largely overestimated in this case. Similar conclusions were drawn by Croteau et al. (2010) for an area also located in eastern Canada.

Related Figures (24)

Table 1 Annual and extreme monthly means for total precipitation (Pio.) and temperature (7 °) for the 1971-2000 period at the Green- wood station
Table 1 Annual and extreme monthly means for total precipitation (Pio.) and temperature (7 °) for the 1971-2000 period at the Green- wood station
Fig. 2. Simplified bedrock geology map based on Keppie (2000) (from Rivard et al. 2012
Fig. 2. Simplified bedrock geology map based on Keppie (2000) (from Rivard et al. 2012
range between 10~’ and 5 x 107° m/s, indicating that a relatively good hydraulic capacity is available throughout the Valley. Field measurements (mainly borehole geo- physics) indicated that movement of groundwater is pre- dominantly controlled by fractures/bedding _ planes. Nonetheless, pumping tests from the database revealed that pumping wells generally behave as if they were in a porous medium, with drawdown following a Theis type curve. Few storage coefficients (S) were reported; they usually range between 10° and 10~? in the bedrock with a median value of 2 x 10~*, suggesting common confined conditions (Todd 1980). A specific study on porosity within the regional study provided values for the Wolfville (28 %) and Blomidon (11 %) formations (see Rivard et al. 2012).
range between 10~’ and 5 x 107° m/s, indicating that a relatively good hydraulic capacity is available throughout the Valley. Field measurements (mainly borehole geo- physics) indicated that movement of groundwater is pre- dominantly controlled by fractures/bedding _ planes. Nonetheless, pumping tests from the database revealed that pumping wells generally behave as if they were in a porous medium, with drawdown following a Theis type curve. Few storage coefficients (S) were reported; they usually range between 10° and 10~? in the bedrock with a median value of 2 x 10~*, suggesting common confined conditions (Todd 1980). A specific study on porosity within the regional study provided values for the Wolfville (28 %) and Blomidon (11 %) formations (see Rivard et al. 2012).
* Obtained from A permeameter testing, B grain-size analysis from Soil Landscapes of Canada (2005), and C pumping tests > O Organic deposits, Lb.v and Ly.b glaciolacustrine deposits, Mb.v, Md, and Ma marine deposits, Ap and At modern fluvial deposits, Mi moderr marine deposits, Go and Gx glaciofluvial deposits, Ca colluvial deposits Table 2 Hydraulic conductivity (K) of the surficial units
* Obtained from A permeameter testing, B grain-size analysis from Soil Landscapes of Canada (2005), and C pumping tests > O Organic deposits, Lb.v and Ly.b glaciolacustrine deposits, Mb.v, Md, and Ma marine deposits, Ap and At modern fluvial deposits, Mi moderr marine deposits, Go and Gx glaciofluvial deposits, Ca colluvial deposits Table 2 Hydraulic conductivity (K) of the surficial units
Table 3 Characteristics of the gauging stations used for the Chapman (1991) base flow separation method *“ This number of years does not necessarily correspond to the period covered, since years were rejected if data were not recorded on a sufficient number of days (usually 300 days per year was taken as the threshold, except if the number of years was small, then 290 was accepted) > Mean total precipitation is 1,180 mm/year. Weather stations are always located at least within 10 km from gauging stations and generally less than 5 km away
Table 3 Characteristics of the gauging stations used for the Chapman (1991) base flow separation method *“ This number of years does not necessarily correspond to the period covered, since years were rejected if data were not recorded on a sufficient number of days (usually 300 days per year was taken as the threshold, except if the number of years was small, then 290 was accepted) > Mean total precipitation is 1,180 mm/year. Weather stations are always located at least within 10 km from gauging stations and generally less than 5 km away
Table 4 Key information related to the five approaches used to estimate recharge in this study. Grey cells represent results obtained from th methods used, whereas white cells represent required inputs of the methods
Table 4 Key information related to the five approaches used to estimate recharge in this study. Grey cells represent results obtained from th methods used, whereas white cells represent required inputs of the methods
rates (J) obtained with the Chapman (1991) method for the seven gauging stations are high, ranging from 347 + 62 to 395 + 94 mm/year, with a weighted average of 364 mm/ year. Despite the fact that watershed areas differ widely (from nearly 9 to 1,020 km’), estimated recharge values only vary by about 25 %. The mean of the six graphical methods for stations 01DC005 and 01DC007 provided even higher values with 460 +91 and 457 + 66 mm/year, respectively. Runoff is very close, on average, to potential recharge estimates when using the Chapman (1991) method, while it is of course considerably ower with the graphical filters, given that baseflow (equivalent to net infiltration) values are much larger. the net recharge W to the bedrock aquifer. It also needs a correction for confined conditions, since percolation to the bedrock aquifer does not occur everywhere. In the FE- FLOW model, net recharge, along with hydraulic conduc- tivities, was used as calibration parameters and they are thus inputs rather than outputs of the groundwater numerical flow model. It is the net recharge to bedrock W that is represented, since the sediment cover in the model could not be well defined and was thin. This sediment cover was thus unable to effectively evacuate subsurface runoff (Rgup). Results from the five approaches are presented in the sec- tion below, taking into account their respective meanings.
rates (J) obtained with the Chapman (1991) method for the seven gauging stations are high, ranging from 347 + 62 to 395 + 94 mm/year, with a weighted average of 364 mm/ year. Despite the fact that watershed areas differ widely (from nearly 9 to 1,020 km’), estimated recharge values only vary by about 25 %. The mean of the six graphical methods for stations 01DC005 and 01DC007 provided even higher values with 460 +91 and 457 + 66 mm/year, respectively. Runoff is very close, on average, to potential recharge estimates when using the Chapman (1991) method, while it is of course considerably ower with the graphical filters, given that baseflow (equivalent to net infiltration) values are much larger. the net recharge W to the bedrock aquifer. It also needs a correction for confined conditions, since percolation to the bedrock aquifer does not occur everywhere. In the FE- FLOW model, net recharge, along with hydraulic conduc- tivities, was used as calibration parameters and they are thus inputs rather than outputs of the groundwater numerical flow model. It is the net recharge to bedrock W that is represented, since the sediment cover in the model could not be well defined and was thin. This sediment cover was thus unable to effectively evacuate subsurface runoff (Rgup). Results from the five approaches are presented in the sec- tion below, taking into account their respective meanings.
obtained with the hydrograph separation, varying from 11 to 195 mm/year, with a weighted average of 82 mm/year (Table 5). Values from the two largest basins (76 mm/year for 01DCO005 and 95 mm/year for 01DC007) are consistent and likely provide the best estimates of flow circulating through the deeper bedrock aquifers. All minimum values were found during August or September. The ratio between the 5-month Chapman method and low-flows varies from 1.2 to 15, mainly depending on the baseflow peaks esti- mated by the numerical filter. The size of the basin does not seem to have a significant influence on the estimate, sug- gesting similar hydrologic conditions for the various investigated watersheds.
obtained with the hydrograph separation, varying from 11 to 195 mm/year, with a weighted average of 82 mm/year (Table 5). Values from the two largest basins (76 mm/year for 01DCO005 and 95 mm/year for 01DC007) are consistent and likely provide the best estimates of flow circulating through the deeper bedrock aquifers. All minimum values were found during August or September. The ratio between the 5-month Chapman method and low-flows varies from 1.2 to 15, mainly depending on the baseflow peaks esti- mated by the numerical filter. The size of the basin does not seem to have a significant influence on the estimate, sug- gesting similar hydrologic conditions for the various investigated watersheds.
Total precipitation at the Greenwood station (in the middle of the Valley) is 1,127 mm/year *“ ETR: real evapotranspiration > “Final” is used when the correction factor for confined conditions has been applied Table 6 Mean values of the various components obtained with the soil moisture balance method for the entire study area and according to areas covering the different geological formations (all in mm/year) over a 20-year period (1980-2000) Hydraulic conductivity (K) and recharge values assigned in the calibrated regional groundwater flow numerical mode are presented in Table 7. Bedrock K values of each unit are usually quite close to the average values obtained from pumping test results. However, K values for the surficia sediments often lie in the upper range of the reported values in Table 2. This likely reflects the fact that (1) K measurements were not necessarily representative of the regional scale, given the heterogeneity of these units and the possibility of fissures and cracks, from which prefer- ential channels develop, and (2) the thinness of surficia sediments covering most of the study area combined with the poor coverage of thickness data often resulted in almost no till cover at all when interpolated in FEFLOW. This is especially evident on South Mountain, which represents nearly half of the study area. This first layer was thus In Fig. 7, it can be observed that values are highest in the center of the Valley, where soils were classified as sandy and well drained in the provincial soil map. Con- versely, values are lowest south of Sheffield Mills, as well as in the area below Coldbrook to Greenwood in the South
Total precipitation at the Greenwood station (in the middle of the Valley) is 1,127 mm/year *“ ETR: real evapotranspiration > “Final” is used when the correction factor for confined conditions has been applied Table 6 Mean values of the various components obtained with the soil moisture balance method for the entire study area and according to areas covering the different geological formations (all in mm/year) over a 20-year period (1980-2000) Hydraulic conductivity (K) and recharge values assigned in the calibrated regional groundwater flow numerical mode are presented in Table 7. Bedrock K values of each unit are usually quite close to the average values obtained from pumping test results. However, K values for the surficia sediments often lie in the upper range of the reported values in Table 2. This likely reflects the fact that (1) K measurements were not necessarily representative of the regional scale, given the heterogeneity of these units and the possibility of fissures and cracks, from which prefer- ential channels develop, and (2) the thinness of surficia sediments covering most of the study area combined with the poor coverage of thickness data often resulted in almost no till cover at all when interpolated in FEFLOW. This is especially evident on South Mountain, which represents nearly half of the study area. This first layer was thus In Fig. 7, it can be observed that values are highest in the center of the Valley, where soils were classified as sandy and well drained in the provincial soil map. Con- versely, values are lowest south of Sheffield Mills, as well as in the area below Coldbrook to Greenwood in the South
*“ Recharge rate over the North Mountain basalts, which are bare or have a very thin surficial cover Table 7 Hydraulic conductivities (K) and recharge rates used ir FEFLOW
*“ Recharge rate over the North Mountain basalts, which are bare or have a very thin surficial cover Table 7 Hydraulic conductivities (K) and recharge rates used ir FEFLOW
Fig. 7 Spatial distribution of annual total infiltration (/) obtained with the soil moisture balance method
Fig. 7 Spatial distribution of annual total infiltration (/) obtained with the soil moisture balance method
Fig. 8 Measured versus modeled heads. The dashed line is the theoretical perfect fit presented as a reference
Fig. 8 Measured versus modeled heads. The dashed line is the theoretical perfect fit presented as a reference
Fig. 9 Spatial distribution of assigned values of annual net recharge to the bedrock (W) in FEFLOW Figure 9 provides the spatial distribution of recharge rates assigned in the FEFLOW model and Table 8 presents values according to the different areas. As values were Mean measured K values had first been assigned in the model, but they were adjusted, keeping K values within measured ranges, until the recharge and other hydrologic components were deemed plausible for this area (recharge below 300 mm/year, surface runoff above 300 mm/year,
Fig. 9 Spatial distribution of assigned values of annual net recharge to the bedrock (W) in FEFLOW Figure 9 provides the spatial distribution of recharge rates assigned in the FEFLOW model and Table 8 presents values according to the different areas. As values were Mean measured K values had first been assigned in the model, but they were adjusted, keeping K values within measured ranges, until the recharge and other hydrologic components were deemed plausible for this area (recharge below 300 mm/year, surface runoff above 300 mm/year,
Table 8 Net bedrock aquifer recharge rates obtained with the numerical model FEFLOW for the entire study area and according to areas covering the different geological units and interflow as high as possible, see below). Minimum values were generally taken, as opposed to those used in FEFLOW, since the HELP model generated larger values of recharge and lower values of interflow (needing very large K contrasts between two layers) than what profes- sional judgments predict. Mean monthly values of total runoff (i.e., the sum of runoff and net infiltration) are compared to streamflows of the 01DC005 gauging station in Fig. 10. The behavior of the curves is similar, suggesting that the model well reproduces the hydrologic system response. Winter values of riverflow are not accurate and thus, no attempt of calibration was sought for the winter period, which still does not depart too much from observations.
Table 8 Net bedrock aquifer recharge rates obtained with the numerical model FEFLOW for the entire study area and according to areas covering the different geological units and interflow as high as possible, see below). Minimum values were generally taken, as opposed to those used in FEFLOW, since the HELP model generated larger values of recharge and lower values of interflow (needing very large K contrasts between two layers) than what profes- sional judgments predict. Mean monthly values of total runoff (i.e., the sum of runoff and net infiltration) are compared to streamflows of the 01DC005 gauging station in Fig. 10. The behavior of the curves is similar, suggesting that the model well reproduces the hydrologic system response. Winter values of riverflow are not accurate and thus, no attempt of calibration was sought for the winter period, which still does not depart too much from observations.
Table 9 Physical parameters of materials used in HELP HELP for the 546 km* 01DC005 catchment. Estimated values are significantly higher than results obtained with the groundwater flow model, mainly because HELP likely underestimates interflow (R.yp). Once water crosses the root zone, it hardly flows out of the cell via interflow and rather tends to reach the water table. In addition, confined conditions are not considered. Also, Environment Canada (personal communication) provided simulated ETR values varying between 511 and 575 mm/year for the study area, suggesting that HELP probably slightly underestimated this parameter (by 10 % or so).
Table 9 Physical parameters of materials used in HELP HELP for the 546 km* 01DC005 catchment. Estimated values are significantly higher than results obtained with the groundwater flow model, mainly because HELP likely underestimates interflow (R.yp). Once water crosses the root zone, it hardly flows out of the cell via interflow and rather tends to reach the water table. In addition, confined conditions are not considered. Also, Environment Canada (personal communication) provided simulated ETR values varying between 511 and 575 mm/year for the study area, suggesting that HELP probably slightly underestimated this parameter (by 10 % or so).
Fig. 10 Validation of total runoff values against streamflows at the 01DC005 gauging station covering 546 km* for the water balance method and the HELP model Table 11 and Fig. 12 present a summary of the mean annual net infiltration J and net recharge W values and their estimated range (including both spatial and temporal var- iability to better assess their potential variability) obtained from the five recharge assessment approaches used in this study. Table 11 also summarizes values of the various hydrological cycle components. According to these results, surface runoff for the entire Annapolis Valley should lie between 300 and 345 mm/year (excluding improbable values from hydrograph separation methods) and real evapotranspiration ETR should be close to 515 mm/year.
Fig. 10 Validation of total runoff values against streamflows at the 01DC005 gauging station covering 546 km* for the water balance method and the HELP model Table 11 and Fig. 12 present a summary of the mean annual net infiltration J and net recharge W values and their estimated range (including both spatial and temporal var- iability to better assess their potential variability) obtained from the five recharge assessment approaches used in this study. Table 11 also summarizes values of the various hydrological cycle components. According to these results, surface runoff for the entire Annapolis Valley should lie between 300 and 345 mm/year (excluding improbable values from hydrograph separation methods) and real evapotranspiration ETR should be close to 515 mm/year.
Fig. 11 Spatial distribution of annual recharge values obtained with HELP over 30 years for the basin represented by gauging station 01DC005 The highest values (>350 mm/year) were provided by hydrograph separation calculated over the entire year and were deemed improbable. When using only the 5 months free of snow and major rain events (June—October), these net infiltration values seem indeed to better estimate the actual contribution of the aquifer (160 mm/year on aver- age) for this geological context. These filters are not physically based and, therefore, do not take into account geological and hydrogeological conditions. They have, however, the non-negligible advantages of being simple and regional. River low-flows provided the lowest values. They likely represent a lower bound of the net aquifer recharge. These methods provide easy-to-obtain watershed
Fig. 11 Spatial distribution of annual recharge values obtained with HELP over 30 years for the basin represented by gauging station 01DC005 The highest values (>350 mm/year) were provided by hydrograph separation calculated over the entire year and were deemed improbable. When using only the 5 months free of snow and major rain events (June—October), these net infiltration values seem indeed to better estimate the actual contribution of the aquifer (160 mm/year on aver- age) for this geological context. These filters are not physically based and, therefore, do not take into account geological and hydrogeological conditions. They have, however, the non-negligible advantages of being simple and regional. River low-flows provided the lowest values. They likely represent a lower bound of the net aquifer recharge. These methods provide easy-to-obtain watershed
Table 10 Various components of the water cycle obtained with HELP for the 01DC005 basin and according to areas covering the differen geological formations (within this basin) over 30 years > “Final” is used when the correction factor for confined conditions has been applied. The hypothesis is that the surplus estimated with the correction factor mainly runs as subsurface runoff
Table 10 Various components of the water cycle obtained with HELP for the 01DC005 basin and according to areas covering the differen geological formations (within this basin) over 30 years > “Final” is used when the correction factor for confined conditions has been applied. The hypothesis is that the surplus estimated with the correction factor mainly runs as subsurface runoff
Table 11 Summary of mean real evapotranspiration (ETR), runoff, total infiltration and net recharge to the bedrock aquifer obtained from the five approaches Mean precipitation is 1,127 mm/year. Bold values represent reliable results * Includes interflow (subsurface runoff) Do A. aero Plas wee pee fe gee eee ere ee l peer ares
Table 11 Summary of mean real evapotranspiration (ETR), runoff, total infiltration and net recharge to the bedrock aquifer obtained from the five approaches Mean precipitation is 1,127 mm/year. Bold values represent reliable results * Includes interflow (subsurface runoff) Do A. aero Plas wee pee fe gee eee ere ee l peer ares
variability, including both spatial and temporal variability (i.e. the larger of the two), to visualize the large variations inherent to each method. Asterisk since FEFLOW was run only in steady-state conditions, only the spatial variability is represented Fig. 12 Summary of mean annual net infiltration (J) and net aquifer recharge to the bedrock aquifer (W) values obtained from the different methods for the entire Valley. The dotted lines represent the most plausible intervals. The first two methods, using 12 months, are presented here only for comparison. Error bars represent total
variability, including both spatial and temporal variability (i.e. the larger of the two), to visualize the large variations inherent to each method. Asterisk since FEFLOW was run only in steady-state conditions, only the spatial variability is represented Fig. 12 Summary of mean annual net infiltration (J) and net aquifer recharge to the bedrock aquifer (W) values obtained from the different methods for the entire Valley. The dotted lines represent the most plausible intervals. The first two methods, using 12 months, are presented here only for comparison. Error bars represent total
(~9,000 '*C years), while this formation is quite perme- able and should, therefore, receive vertical recharge. FE FLOW and HELP models) cannot be made since cell characteristics greatly vary, both due to their difference in size and different parameters taken into account. None- the ess, recharge rates for the various geological formations were estimated. They provided values varying between 60 and 170 mm/year for the less permeable formations (South Mountain Batholith and Undifferentiated Lower Paleozoic roc ks) and 90 to 255 mm/year for the most permeable formations (North Mountain basalts, Blomidon, and Wolfville), depending on the method. As expected, FE- FLOW generally provided the lowest values and HELP the highest values.
(~9,000 '*C years), while this formation is quite perme- able and should, therefore, receive vertical recharge. FE FLOW and HELP models) cannot be made since cell characteristics greatly vary, both due to their difference in size and different parameters taken into account. None- the ess, recharge rates for the various geological formations were estimated. They provided values varying between 60 and 170 mm/year for the less permeable formations (South Mountain Batholith and Undifferentiated Lower Paleozoic roc ks) and 90 to 255 mm/year for the most permeable formations (North Mountain basalts, Blomidon, and Wolfville), depending on the method. As expected, FE- FLOW generally provided the lowest values and HELP the highest values.
Related topics:
Earth and Environmental Sciences
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved