a concept for estimating depths to the for catchment scale...

21
A concept for estimating depths to the redox interface for catchment scale nitrate modelling in a till area Anne Lausten Hansen ([email protected]) 1,2 Christensen BSB 3 , Ernstsen V 1 , He X 1 and Refsgaard JC 1 (1) Geological Survey of Denmark and Greenland (2) Department of Geosciences and Natural Resource Management, University of Copenhagen (3) Rambøll

Upload: others

Post on 24-May-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

A concept for estimating depths to the redox interface for catchment scale 

nitrate modelling in a till area Anne Lausten Hansen ([email protected])1,2

Christensen BSB3, Ernstsen V1, He X1 and Refsgaard JC1

(1) Geological Survey of Denmark and Greenland(2) Department of Geosciences and Natural Resource Management, University of Copenhagen (3) Rambøll

Page 2: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

• Nitrate can be naturally transformed by reduced compounds(OM, Fe+2, pyrite) in the sediments

• Transition from oxic to reduced conditions = redox interface

• Spatial variation in the redox interface and in water flow paths leads to nitrate sensitive and nitrate roboust areas

• Important to know the location of the redox interface to delineate these areas

Introduction

Nitrate sensitive area Nitrate roboust area

Introduction Background Methodology Application Evaluation Future

Page 3: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Introduction• Location of the redox interface in till areas varies several meters within short distances

• The interface can only be determined by drilling boreholes => Limited data 

Large uncertainty on the location of the redox interface

• Interesting to develop methodologies to inferthe location of the redox interface from othervariables

Introduction Background Methodology Application Evaluation Future

Objective of this study

Page 4: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Redox interface development ‐ Hypothesis‐

• The reduced compounds (redox capacity) in the sediments is depleted by oxygen and nitrate

• The present location of the redox interface is the result of the cumulative flux of oxygen in recharging groundwater since the onset of Holocene (11.700 years)

• Development of the interface in parts of the unsaturated zone can have happened fast due to oxygen diffusion in the air phase. In a clay till, however, this is only important in the root zone

Introduction Background Methodology Application Evaluation Future

Page 5: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Redox interface concept• Key principle: estimate the spatial pattern of the redox interface

from variability in groundwater recharge and sediment redoxcapacity

• Redox equation: the redox depth in grid i is estimated as:

Redox depthi = fluxi ∙ f + min. redox depth 

flux: recharge flux estimated with hydrological model f: redox interface migration constant (m over 11.700 pr mm yearly recharge)  min. redox depth: Upper part of UZ where redox capacity have been depleted fast 

due to air phase diffusion

Additional parameters:  Maximum redox depth Lower redox depth in riparian lowlands

Introduction Background Methodology Application Evaluation Future

Dependent on the sediment redox capacity !

Page 6: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Step 1:Extraction of recharge flux from hydrological model (no drainage and pumping)

Step 2:Difference in redox capacity betweensediment types applied to recharge map

Step 3:Apply redox equation, define f for mainsediment type

Step 4:Run nitrate model with estimated redoxinterface => Simulated nitrate arrival (% of nitrate input, NAP) at catchment outlet

Step 5:Compare simulated and observed NAPIf sim >< obs => new constant f and min. redox depth

Redox interface concept

Introduction Background Methodology Application Evaluation Future

Page 7: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Application in Norsminde fjord catcment

Introduction Background Methodology Application Evaluation Future

Topography Soil type Redox depthobservations

Page 8: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Models• Geological model

– 11 hydrogeological units – Based on borehole data from Jupiter and 

geophysical data from Mini‐SkyTEM

• Hydrological model– MIKE SHE/MIKE 11– All hydrological processes– Grid scale 100x100 m

• Nitrate model– Particle tracking (MIKE SHE)

Introduction Background Methodology Application Evaluation Future

Page 9: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Nitrate model ‐ particle tracking

• Nitrate input: Daily N leaching from rootzone

N balance method combined with Daisy simulations  (Thirup (2013), available at www.nitrat.dk)

• Redox interface implemented as registration zone => particle registreted if crossing interface

• Nitrate arrival: particles arriving in fjord without crossing redox interface

• The model is run 4 years with N input (2000‐2003) and then additional 4 yearsto get all nitrate out (flow recycled)

Distribution of particles at different sim. time(N added first 4 years)

Introduction Background Methodology Application Evaluation Future

Page 10: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Calibration target‐ Nitrate arrival percentage (NAP) to Norsminde fjord ‐

Introduction Background Methodology Application Evaluation Future

41 – 49 % of the nitrate leaching arrives in Norsminde fjord

Input period  1998‐2005 1998‐2004 2000‐2003 2000‐2003 2000 ‐ 2005 2000 ‐ 2004Obs period 1998‐2005 1999 ‐ 2005 2000‐2003 2001‐2004 2000 ‐ 2005 2001 ‐ 2005Avg. N leaching input [t/yr] 365 368 281 281 279 267Avg. obs N flux to fjord [t/yr] 157 151 127 135 131 130NAP [N leaching/N flux] 0.43 0.41 0.45 0.48 0.47 0.49

Page 11: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Redox scenarios and calibration• Redox scenarios (based on sensitivity analysis)

– Scenario 1:  Recharge flux layer 2 (3 ‐ 4 m.b.s) Redox depth in riparian lowlands 1.5 m

– Scenario 2:  Recharge flux layer 1 (0 ‐ 3 m.b.s)Redox depth in riparian lowlands 1.5 m

– Scenario 3:  Recharge flux layer 2 (3 ‐ 4 m.b.s) No riparian lowlands

• Calibration– All 3 scenarios was calibrated to NAP = 45%

Introduction Background Methodology Application Evaluation Future

Scenario Constant f Min. redox depthScenario 1 0.025 2.65Scenario 2 0.0155 1.5Scenario 3 0.025 2.5

Calibrated parameter valuesNorsminde redox data (clay till)Avg. redox capacity: 418 meq‐e/kgO2 conc.: 11.4 mg/l (10oC)=> Constant f = 0.025

Page 12: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Introduction Background Methodology Application Evaluation Future

Redox interface and reduction maps

Page 13: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Estimated versus observed redox depths‐ point scale ‐

Introduction Background Methodology Application Evaluation Future

Page 14: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Estimated versus observed redox depths‐ catchment scale ‐

Introduction Background Methodology Application Evaluation Future

Page 15: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Evaluation of Results

• The model is able to simulate observed nitrate arrival (NAP)to Norsminde fjord

• All 3 scenarios can be cailbrated to NAP = 45% => equifinality

• Redox depth observations not sufficient to choose between scenarios

• Cumulative distribution of redox depths close to observed

• Site‐specific redox depths is not well estimated

• Results okay on cathment scale, but not on small scale

Introduction Background Methodology Application Evaluation Future

Page 16: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

• Recharge flux– Constant flux– Only vertical component of flux

• Migration constant f– Uniform migration constant f within sediment type– Variation in sediment type with depth not included

• Scale issue (Model grid scale 100x100 m)– Affects estimated redox depths due to averaging– Affects compariosn of estimated vs. obseved redox depths

• Nitrate data– N leaching– N flux to Norsminde fjord

• Geological and hydrological model– Flow paths correct ?

Factors affecting the results

Introduction Background Methodology Application Evaluation Future

Norsminde dataRedox capacity (clay till)Avgerage: 418 meq‐e/kgSt.dev.: 150 meq‐e/kg

Page 17: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Conclusions

• The concept is capable of estimating the general location of the redox interface, but not at grid scale

• The model is therefore not able to accurately simulate nitrate reduction at grid scale

• The uncertainty on the reduction potential maps needs to be evaluated

Introduction Background Methodology Application Evaluation Future

Page 18: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Work in progress‐ Application of redox concept on 20 geological models ‐

Introduction Background Methodology Application Evaluation Future

Uncertainty on nitrate reduction at different aggregation scales

Page 19: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Thank you for your attention!

Page 20: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Redox depth observations

• Log‐normal distributed (p‐value 0.8) with a mean of ln(redox) = 1.6 m => redox = 4.7 m

• No correlation to other variables (elevation, distance to stream, min. water table)

• A varioagram analysis showed spatial correlation with a correlation length of 289 m

Only resolved by a few clusters ofboreholes and not the entire data set=> not representative for whole area

Introduction Background Methodology Application Evaluation Future

Page 21: A concept for estimating depths to the for catchment scale …nitrat.dk/xpdf/dwf_jan2014_alha.pdf · A concept for estimating depths to the redox interface for catchment scale nitrate

Redox interface development ‐Mass balance example ‐

Borehole profile from Lillebæk (LOOP4)Meq e‐/kg : milli‐electron‐equivalentsper kg sedimentData from Ernstsen (2013), available at www.nitrat.dk

Introduction Background Methodology Application Evaluation Future

1. Redox capacity (reactive):  450 meq e‐/kg2. Bulk density:  1590 kg/m3

3. Electron use (O2 reduction): 4 e‐/mole= 0.125 meq e‐/mg

4. O2 conc. (10 oC): 11.4 mg/l5. Recharge rate: 273 mm/yr

Migration pr. year (3*4*5)/(1*2): 5.4e‐4 m/yrTotal migration (11.700 years): 6.4 m (below root zone)

Migration pr. mm yearly recharge: 0.023 m/(mm*yr‐1)

Migration constant f‐ Independent on recharge flux‐ Very dependent on redox capacity