Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


The Use of MODFLOW in Scientific Literature

Jerome Molenat and Chantal Gascuel-Odoux (2002)

Laura Johnstone

on 22 October 2012

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of The Use of MODFLOW in Scientific Literature

Jerome Molenat and Chantal Gascuel-Odoux (2002) Modelling flow and nitrate transport in groundwater for the prediction of water travel times and of consequences
of land use evolution on water quality How is groundwater flowing? Research Questions and Model Purpose Site Conditions Model Design & Parameters Calibration Validation Spatial Patterns of Nitrate and Sulfate Patterns & Hypotheses How is nitrate being transported in the groundwater? What is the effect of changes in nitrate leaching into groundwater - due to changing agricultural practices - on the nitrate concentration in stream water? What is the water residence time in
the groundwater? Catchment Scale The model was used as a tool to predict and provide insight into the efficiency and consequences of land use changes and alternative agricultural practices used to reduce water pollution. Guidance needed to determine the best methods to put in place. Put in place new land use and agricultural management strategies to reduce nutrient load in the water. Background In French Brittany, surface water provides most of water consumed One of the most productive agricultural regions in Europe An increase in nitrate concentrations in stream water accompanied an increase in agriculture 80% of surface water exceeds European regulation levels for nitrate in drinking water Parameters Recharge = 2mm/day In scenarios 3-6 nitrate recharge varied spatially between 50-150 mg/l.
In scenarios 1 & 2 nitrate recharge was decreased to 80 and 60 mg/l resp. 2D, steady state model representing a vertical section perpendicular to the groundwater surface, taken from the hillslope Four Model Units with variable thickness:
1. The Plough Layer
2. The Soil (loamy silt mixture of weathering products and Quaternary eolian
3. Weathered Shale
4. Fissured Shale Homogeneous and isotropic within formations Topography:
The base of the shale aquifer was set at 70m and hillslope G reached an elevation of just under 20m.
Slopes in the area do not exceed 10%, the
elevation of which ranges 98-140m a.s.l. Grid size: Variable but never exceeded 5m Boundaries:
Assumptions: no lateral flow at the top of the hillslope and at the base of the shale.
Constant head in the stream water
Uniform recharge along the hillslope. The authors calibrated hydraulic conductivity each unit by trial and error. They fitted the modelled hydraulic heads to the observed heads while keeping the hydraulic conductivity values in
the initial estimated known range. The finite-difference code MODFLOW was used to simulate the distribution of hydraulic head within the groundwater. Nitrate transport was described by the convection – dispersion equation solved using MT3D. MODPATH was also used to analyse flow paths and travel times within the groundwater 6 different scenarios No validation? "...the travel times computed by the model would have to be validated further. The assessment of
the model on hydraulic heads and nitrate concentrations must be considered as the first step. Direct estimation
of travel times from water age dating methods based on tracers could be a way to go further in the validation." Scenarios assumed to represent evolution of Brittany’s agriculture towards better management practices Solution 1: Reduce nitrate leaching from all fields.
In scenarios 1 & 2 a uniform decrease of nitrate concentration in the recharge from 100 mg/l to 80 mg/l and 60 mg/l respectively. Solution 2: Nitrate concentration in recharge is spatially distributed (Based on the idea that, for a given average nitrate leaching rate at catchment scale, an optimum spatial distribution of nitrate leaching rate would allow a minimum nitrate concentration in stream water). This solution was used in scenarios 3-6 Solutions for the restoration of water quality Field Conditions Total Porosity & Hydraulic conductivity Soil: 'Realised from undisturbed soil blocks', giving K and water retention curve Weathered Shale: Pumping test on piezometers Fractured/fissured Shale (regarded as homogenous porous aquifer): Pumping tests on deep piezometers. Nitrate, Sulfate & Chloride Fortnightly ground- and stream water samples taken and analysed by ion exchange chromatography Weathered shale groundwater appeared to be major hydrologic storage of nitrate controlling the export of nitrate to the stream. Nitrate (white circles) concentrations and sulfate concentrations displayed uphill and downhill gradients respectively. *Shale nitrate, sulfate and chloride concentrations were 2.8mg/l, 20.8mg/l and 18.0mg/l respectively. Weathered Shale Groundwater Chemistry Why the strong spatial pattern (*high sulfate, low nitrate), in the weathered shale groundwater chemistry? *remember me! Dilution with water containing a low level of nitrate. Low nitrate concentrations? Attributed to autotrophic denitrification together with the oxidation of pyrite (FeS2), which is present in the shale, and the consequent production of sulfate. Both the nitrate and sulfate distributions observed in the upper layer of the weathered shale groundwater explained by upward flows of deep denitrified groundwater to the weathered shale aquifer. Effective porosity in fissured shale taken as equal to total porosity.
For loamy silt and clay mediums (weathered shale, soil and plough layer), effective porosity considered as equal to specific yield. Convection assumed to be main process in nitrate transport, so dispersion and molecular
diffusion were neglected in the model In soil

"Previous studies on heterotrophic denitrification in hydromorphic soils of the KC revealed that rates were very high, such that nitrate was completely reduced within a few hours, especially when initial concentrations were high. However, a first-order reaction with a very small half-life seemed to be appropriate to describe the fast nitrate reduction observed in the field. Thus, the half-life of heterotrophic denitrification was fixed at 0.5 day." In Shale

"Autotrophic denitrification assumed to occur in pyrite-rich layers. Deep drillings revealed that, in hillslope G, pyrite was present in all the shale and in the weathered
shale beyond 20 m deep (Figure 2b). The half-life was fixed at 5 days, in accordance with field tracer tests performed in the shale aquifer of Naizin by Pauwelset al. (1998) showing that autotrophic denitrification could be described by a first-order equation with a half-life ranging from 2.1 to 7.9 days." Results How is groundwater flowing? What is the water residence time in the
groundwater? How is nitrate being transported in the groundwater? What is the effect of changes in nitrate leaching into groundwater - due to a change in agricultural practices - on the nitrate concentration in stream water? Model reproduces observed heads well Discrepancy between simulated and observed heads increases uphill Mean error (ME) = 0.8m
Mean absolute error (MAE) = 1.10 m. Nitrate concentration ME 3.8 mg/l (estimated)
MAE = 7.9 mg/l (estimated) The model represented the decrease of nitrate concentration along the hillslope well. Nitrate conc. simulated in PG2 (nearest stream) had higher than the average annual conc. Deviation was 24mg/l - 61% error, compared with a max. of 6mg/l - 5% error for all others 3 main groundwater flow paths
1. Reaches deeper shale aquifer
2. Midslope - Through weathered shale aquifer and soil
3. Through plough and soil layer straight to stream. Travel times for a 'non-reactive particle' varied from few days, entering at very bottom of slope to 3 years, entering at top. Any particle entering on top 60% of slope had residence time of at least 1 year. Greatest changes of stream water nitrate conc. obtained following an overall and uniform reduction of nitrate leaching (scenarios 1 and 2). 2 zone distribution - (scenarios 3 and 4), caused only small changes in nitrate conc., not greater than 9%. 3 zone distribution - Changes more pronounced than 2 zones. Higher nitrate conc's in the midslope caused increase of nitrate conc's in the stream (scenario 5), whereas higher nitrate conc's in the upper zone and in the bottom land caused a decrease of nearly 20% (scenario 6). Nitrate conc. equilibrium reached after around 1200 days. 3 patterns of nitrate evolution emerged:
1. A progressive decrease beginning in first 100 days (scenarios 1 & 2)
2. Initial increase followed by a decrease (scenarios 4 & 6).
3. Initial decrease followed by an increase (scenarios 3 & 5). Conclusions The model incorporated some simplifications - may explain discrepancy between simulated
and observed hydraulic heads in summit of the hillslope. Successful simulation of the field nitrate transport conditions model, showing decrease of nitrate conc. in the weathered shale aquifer along the hillslope. Hypothesis Confirmed - Simulations show that decrease resulted from upward flows from deep denitrified groundwater. Yet simulated nitrate conc. in piezometer
PG2 (nearest stream) was greater than that observed. Is difference due to denitrification within the upper part of weathered shale? Not represented in the model. But did denitrification occur in all the
weathered shale aquifer? - Would have observed a downhill decrease of nitrate conc.
from PG6 to PG2. Rather, decrease restricted to piezometers PG2 and PG3. Deviation between
simulation and observation raises question of difference between the scale of measurement and the scale of modelling. Difficult to compare a point value with a cell with a size of a few metres. Rather than comparing point values with grid values, spatial patterns of state variables could be used to assess the accuracy of models. Going by those criteria, model reproduces the main characteristic of the groundwater chemistry -
the strong spatial pattern of nitrate concentration. Nitrate Concentrations It appears, from simulations, that one half of the nitrate removal is caused by autotrophic
denitrification in the pyrite-rich layers, the remainder being attributed to heterotrophic denitrification. Residence Times Few days to many years - important implications for water quality restoration. Any reduction in the nitrate leaching to groundwater will cause an immediate but gradual decrease in stream nitrate concentration. In highly contaminated catchments, persistent high nitrate conc. expected even if remediation measures are undertaken (longest travel time estimated to be more than 3 years)... ...Not taking into account the immobile water (water in micro-porosity or in dead-end pores in shale). Simulations carried out under
winter recharge conditions... ...Highest hydraulic gradients in the year... ...Fastest water velocity... ...Groundwater travel times from model must be regarded as underestimates. Also did not consider nitrate travel time from
soil surface to water table. Scenario 6 - takes efficient advantage of denitrification processes, average nitrate leaching was 165 kgN /ha/yr, Scenario 1/2 - lowest average nitrate leaching rates In scenario 6, highest nitrate recharge rates are added to the most upslope and downslope zones, which are the starting points of flows travelling through denitrifying layers. A radical decrease of nitrate concentration in stream water would require a ‘global’ and large reduction of average nitrate leaching, as demonstrated by the result of scenario 2. Which scenario works best? Next Steps? 1. Consider a transient model with variable groundwater recharge. 2. ?...
Full transcript