The inflows and outflows of river leakage appear to remain fairly consistent over time

By plotting the residuals on a map of the parent model domain, we were able to assess the distribution of wet and dry wells in our model. Blue colors indicated an excess of water in certain wells, red and orange colors indicated wells that were more “dry,” and green colors represent a satisfactory residual value. A satisfactory residual value means that the difference between the observed and simulated equivalent heads were small, indicating good model performance. The black square on the residual map is a reference point for the area of the child model domain’s boundary.The outputs of the parent model are data of groundwater fluxes, which includes the magnitude of volumetric flow through each model cell and the direction. The groundwater flux was calculated at each cell in every layer of the model for each stress period . The hydraulic head, or groundwater elevation above sea level in meters, was plotted as orange contours with the magnitude and direction of groundwater flux for Layers 1, 6 and 9 with black arrows . These three layers were chosen out of all nine layers because they were deemed representative of the top, middle, and bottom of the model’s layers. The length of each arrow represents the magnitude of groundwater flux, with the arrow pointing in the direction of flow. The map also includes a black square that serves as a reference point of the child model area, and the flow that occurs at and around the child model’s boundary at each layer. The maps show that the general flow of groundwater in the parent model is towards the cone of depression and the pumping wells in that area, with water flowing away from the Sacramento River near the recharge sites in the child model area.The results of Scenario 1, which was run using the child model, represents the baseline model run performance and outputs,blueberry pots using the original and unaltered deep percolation data from Davids Engineering. For this baseline scenario, we assessed the model performance using a similar analysis to the parent model performance.

The one-to-one plot for the child model is a plot of the observed and simulated equivalent heads with a gray line indicating the line of equal value . The colors of the dots represent the binned residual head data.The binned residual heads were also plotted on a map of the child model domain, indicating the location of the four groundwater monitoring wells used to collect continuous and periodic groundwater elevation data during the recharge studies at each ag-MAR field site . The location and shape of the field sites are outlined in black, and the four colored dots show the location of the groundwater monitoring wells and their average residual simulated head value. Yellow colors indicate “dry” wells , green colors represent good model performance , and blue colors indicate “wet” wells .The child model outputs of Scenario 1 include the magnitude and flow direction of volumetric groundwater fluxes in each model cell, calculated using MODFLOW’s groundwater flow equation at each stress period in the model run; the child model is transient and has 451 stress periods, or days, that it simulated in its model run. Head contours are plotted as orange lines and flow magnitude and direction plotted as black arrows illustrate the direction and magnitude of groundwater flow . Groundwater generally flows southwest, with water flowing away from the Sacramento River in the area near the recharge sites. Only the first layer was plotted, since this is the only layer of interest due to it being the layer of the child model in which the RCH package defines recharge to occur.A more in-depth analysis of the model performance was conducted to understand how the observed and simulated equivalent heads change over time at each of the four groundwater monitoring wells in the child model domain. Hydrographs of the observed and simulated equivalents were plotted for each groundwater monitoring well; the head, or groundwater elevation ASL was plotted on the primary y-axis, while recharge rates was plotted on the secondary y-axis as a bar plot .

Hydrographs and recharge rates were plotted together for each well to illustrate the relationship between groundwater elevation and recharge rate over time. In half of the groundwater monitoring wells we observed simulated equivalent values greater than the observed values, and in the other half of the wells we observed simulated equivalent heads less than the observed values. The difference between the observed and simulated equivalent head values were the greatest for the Field_15 well, with a 12-14 m difference, while the difference between the observed and simulated values for the other three wells were significantly less, with a 2-6 m difference. This discrepancy may arise from the spatial variation in where the groundwater monitoring wells are located . The Field_2_2019, Well_2b, and Well_34 monitoring wells are all located near each other, near the Sacramento River. In contrast, the Field_15 well is located the farthest away from the Sacramento River and the other three wells. Recharge rates varied over time among the four groundwater monitoring wells. For Field_2_2019 and Field_15, recharge gradually increased in mid- to late September of 2019 and remained constant throughout the remainder of the flooding period in October of 2019. For Well_2b and Well_34, recharge rates increased rapidly in September of 2020, remained fairly constant through October of 2020, and slowly decreased during November of 2020 .Additional child model outputs from Scenario 1 include the water budget components that flow into and out of the model’s cells, calculated during each stress period of the model run. The water budget components plotted include storage , river leakage , recharge , storage and river leakage . A stacked bar chart was made to represent the magnitude of the rates of each water budget component over time . Water that flows into the model cells from storage, river leakage, and recharge are plotted in light green, light orange, and blue, respectively. These inflows are specified as positive rates on the y-axis.

Water that flows out of the model cells and are transferred to storage and river leakage are plotted in dark green and dark orange, respectively. These outflows are specified as negative rates on the y-axis. The rates of each component were plotted to see the variations over time, such as seasonal differences in recharge or storage. The child model results of Scenario 2 represent increased rates of deep percolation during the same recharge periods as Scenario 1, where deep percolation values were increased by one order of magnitude. Similar to Scenario 1, we analyzed the hydrographs of observed and simulated equivalent heads over time at each of the four groundwater monitoring wells . Groundwater elevation levels were plotted on the primary y-axis, and recharge rates were plotted as a bar chart on the secondary y-axis to compare the observed and simulated equivalent heads with recharge rates over the course of each recharge period. Both simulated equivalent head values were plotted from Scenario 1 and Scenario 2 to display the hydrologic difference between the two scenarios. Overall, the hydrographs of the simulated equivalent heads for Scenario 2 are consistently higher than those of Scenario 1. Recharge rates varied over time among the four groundwater monitoring wells for Scenario 2. For Field_2_2019 and Field_15, recharge gradually increased in mid- to late September of 2019 and remained constant throughout the remainder of the flooding period in October of 2019. For Well_2b and Well_34, recharge rates increased rapidly in September of 2020, remained fairly constant through October of 2020,best indoor plant pots and slowly decreased during November of 2020 .The water budget components extracted from the child model results of Scenario 2 were represented with a stacked bar chart . Each of the components plotted are the same as the components plotted for Scenario 1. The water budget components plotted are storage , river leakage , recharge , storage and river leakage . The main difference is that the scale of the rates of the components is several times larger than that of Scenario 1. Recharge rates are largest during the flooding periods at the ag-MAR field sites. Inflows to the model cells from storage are also greatest during the periods of high recharge rates, and outflows from the model cells to storage start to increase after a few weeks of recharge.The results of Scenario 3 represent a repetition of the 2019 recharge program for ten consecutive years.

Because of the large difference in time discretization between Scenarios 1 and 2 and Scenario 3, the analysis of the model performance of Scenario 3 was conducted individually without an initial comparison with the first two scenarios. Hydrographs of the observed and simulated equivalent head values were plotted for each groundwater monitoring well on the primary y-axis, while recharge rates were plotted on the secondary y-axis . The groundwater elevation and recharge rates were plotted together to demonstrate their relationship over the course of the ten years. The simulated head values were plotted alone for this ten-year model run because there are no projected groundwater elevation data from the year 2020 to 2029. The pattern of the hydrograph is repetitive, as each consecutive year is a replication of the previous one. The same can be said for the bar chart of the recharge rates. As recharge rates peak, the simulated heads also reach their peak.The water budget components that were an output of the child model for Scenario 3 are plotted as a stacked bar chart . The components of interest that were plotted as rates either into or out of the child model’s groundwater system included storage , river leakage , recharge , storage and river leakage . The rates of these inflows and outflows to the groundwater system span the ten years that the model was run for Scenario 3. The graph follows a consistent pattern, for each component, as the whole ten years is a replication of the first year . There is no visible increase in any of the water budget components’ rates over time, as the pattern is very consistent.From analyzing the model performance of the parent model through the one-to-one plot of observed and simulated equivalent heads , it is evident that most of the data are a good fit to the one-to-one line, with the exception of the cluster of red dots in the lower left area of the plot. This group of heads had a residual greater than -100 m, meaning the difference between the observed and residual heads at these groundwater monitoring wells were greater than -100 m. Other than this cluster of red residuals, the binned residual head data indicated that the majority of the values of the observed and simulated equivalent heads were fairly close, with a difference of less than -100 m to 30 m. The overall shape of the distribution of dots on the one-to-one plot suggest that we were mostly underestimating the heads in certain areas, resulting in a more conservative estimation of heads in the parent model domain. The residual map of the parent model shows the distribution of groundwater monitoring wells and how dry or wet they are according to the color of the well points . The northern half of the parent model domain showed good to fairly good model performance, with the green and yellow colored residual points indicating a small difference between the observed and simulated equivalent values of head. Within the black square, the area of the child model domain, most of the groundwater monitoring wells indicate an underestimation of heads . The areas where the residuals were smallest were areas near surface water features along the eastern boundary of the parent model domain, such as Butte Creek, Butte Slough and the West Borrow Ditch, as well as some parts of the Sacramento River in the domain area. The cluster of red dots that were shown in the one-to-one plot are seen in the middle to southern half of the parent model domain . The red dots indicate an area of low groundwater elevation, forming a cone of depression. The cone of depression shown by the parent model results is not a surprising discovery.