PHD RESEARCH PROGRESS Assessing of the impact of on-site sanitation on man-induced recharge and groundwater quality in the Thiaroye aquifer of , . DIAW Mor Talla, UCAD, Senegal

Third international AfriWatSan Stakeholders Workshop Kisumu, Kenya Main Assessing of chemical and microbiological faecal contamination Objective and hydrodynanism processes related to the impact of on-site sanitation

Assess the contribution of on-site sanitation to groundwater-level 1 dynamics (hydrodynamic regional model) ;

Assess the pathways and processes of chemical and microbiological 2 contamination (transport models in miniature, statistics and multi- criteria models) ;

Assess variations in faecal contamination of groundwater using

3 hydrochemical and microbriological indicators (NO3, PO4, E.coli) and their relationship to human health outcomes. 2 Thiaroye aquifer (Dakar) & a low-income on-site sanitation area

Built

Pambal built

Malika

Sanghalkm Wakhinane Nimzatt Ndiareme Limamoulaye Yeumbeul Nord Keur Massar occupying

Sam Notoire Golf Sud camberene Medina Gounass Yeumbeul Sud Djidah Thiaroye Kaw Nord Thiaroye Gare Pikine Ouest Pikine Est about 80% Guinaw Rail Nord Diamaguene Sicap Mbao Mbao Patte d'Oie Guinaw Rail SudThiaroye Diaksao Ngor Dalifort Thiaroye Sur Mer Ouest

Rufisque Nord of the total Sicap Liberté Dieupeul-Derkle

Mermoz-Sacré Coeur HLM SEBIKOTANE Boundaries Hann-Bel Air area of the Senegal Fann-Point E- Amitié Gueule Tapée, Fass-Colobane Acquifer localisation DAKAR Medina urban area Marly substratum Plateau 3 Tanma Lake Quaternary Sands Aquifer 3 1.5 0 Km AFRIWATSAN Í 4 2 0 4 Clay and volcanic rock r Km AFRIWATSAN Í - Aquifer used to account for 10% of the Dakar ‘ S water supply Shallow aquifer, Quaternary sands High contamination (NO3, Cl) Supply borehole stops (2008) aquifer, Extreme west of Senegal

On-site sanitation Sewered sanitation on-site sanitation Cité Keur Salam "

Malika (septic

Wakhinane Nimzatt Sanghalkm Ndiareme Limamoulaye Yeumbeul Nord Keur Massar tanks)

Sam Notoire Medina Gounass camberene Golf Sud Yeumbeul Sud Djidah Thiaroye Kaw

Parcelles Assainies pikine Nord Thiaroye Gare Pikine Ouest Pikine Est Guinaw Rail Nord Diamaguene Sicap Mbao Thiaroye Diaksao Mbao Yoff Patte d'Oie Guinaw Rail Sud Ngor Dalifort Thiaroye Sur Mer Rufisque Ouest Grand Yoff Rufisque Nord Sicap Liberté Ouakam Dieupeul-Derkle

HLM Coverage Mermoz-Sacré Coeur Biscuiterie

Grand Dakar Hann-Bel Air rate (90 - Fann-Point E- Amitié Gueule Tapée, Fass-Colobane Medina 95%) Plateau 3 0 3 Km AFRIWATSAN Í Hydrohemical and microbiological results

53 % Large 12 wells > number of sites 50 mg/L 10 affected by 8

faecal (meq/L)

- 3 6 contami NO nation Max 710 4 infuced

mg/L the marine marine the

by

2 marine under intrusio

60 % 0 n

Water Water influence

wells > 0 2 4 6 8 10 12 14 16 18 20 250 Cl-(meq/L) mg/L - Increase in contamination caused by single injection of septic tanks

- Recharge of rainfall origin Lowers contamination

August 2016 April 2017 800 Linear (August 2016) Linear (April 2017) 700

70 % 600

wells > - 500 1 CFU /100 NO3 400 mL 300 200

100

0 4 On-site sanitation key input (Sanitation facilities)

 Advanced spatial tool for determining the number of septic tank /pit

 Census of the sanitized district communities “Actualisation du Plan Directeur d'Assainissement Liquide de Dakar 2025 »  Object-oriented classification of the Quickbird satellite image (2016)

 3 Auxilary high resolution satellite image

Data type acquisition Source Résolution

Satellite image 2003 QuickBird 0.70 (pan-shapping… Satellite image 2009 Orbview 1.60 HR satellite image 2012 Geoeye 0.60

SEGMENTATION

CLASSIFICATION

VALIDATION On-site sanitation input (Sanitation facilities)

 Advanced spatial tool for determining the number of septic tank/pit

On-site sanitation Sewered sanitation

Cité Keur Salam "

Malika

Wakhinane Nimzatt Sanghalkm Ndiareme Limamoulaye Yeumbeul Nord Keur Massar

Sam Notoire Medina Gounass camberene Golf Sud Yeumbeul Sud Djidah Thiaroye Kaw

Parcelles Assainies pikine Nord Thiaroye Gare Pikine Ouest Pikine Est Guinaw Rail Nord Diamaguene Sicap Mbao Thiaroye Diaksao Mbao Yoff Patte d'Oie Guinaw Rail Sud Ngor Dalifort Thiaroye Sur Mer Rufisque Ouest Grand Yoff

Rufisque Nord Sicap Liberté Ouakam Dieupeul-Derkle

Mermoz-Sacré Coeur BiscuiterieHLM - West zone (Sewered sanitation) or 74 Km2 Grand Dakar Hann-Bel Air

Fann-Point E- Amitié - East zone (On-site sanitation) or 99 Km2 Gueule Tapée, Fass-Colobane

Medina

Plateau 3 0 3 Km AFRIWATSAN Í Fig 1: Distribution of sanitation type (DIAW,2017) On-site sanitation key input (Sanitation facilities)

 Advanced spatial tool for determining the number of septic tank/pit

 A 100x100 m grid at the scale of the aquifer  In densely built areas (Spectral and textural characteristics)/size of house  In low-built areas (Visual interpretation of Geoeye image)  10 grids of 300 x 200 m for validation

Density of septic tank/ Validation grid pit latrine per 10000 m 2 01 - 10 11 - 16 17 - 23 24 - 30 31 - 37 Malika 38 - 44 45 - 50

51 - 55 7 Sanghalkm Wakhinane Nimzatt 56 - 61 Ndiareme Limamoulaye Yeumbeul Nord Keur Massar 62 - 70 350 Sam Notoire Y = 0.9914x + 6.0714 3Medina Gounass 300 camberene 2 Golf Sud 4 R² = 0.9977 Yeumbeul Sud Djidah Thiaroye Kaw 6 250 pikine Nord Parcelles Assainies Diamaguene Sicap Mbao Pikine Est Thiaroye Gare 200 1 Pikine Ouest Guinaw Rail Nord 150 Patte d'Oie Thiaroye Diaksao Mbao

through through field work Guinaw Rail Sud

100 Thiaroye Sur Mer

Dalifort 5 Number Number of septic tank obtained 50 Rufisque Ouest Hann-Bel Air 0 Rufisque Nord 0 40 80 120 160 200 240 280 320 360 3 0 3 Number of septic tank determined via Quickbird image Km AFRIWATSAN Í On-site sanitation input (Sanitation facilities)

 Advanced spatial tool for determining the number of septic tank

Density of septic tank/ 2 pit latrine per 10000 m Number of septic tank/pit ≈ 253116 Deni, Bayakh, Mbaouane… High septic 01 - 10 Yoff Grand yoff tank/pit 11 - 16 Ouakam 17 - 23 Thiaroye Sur Mer density in Dalifort 24 - 30 Hann-Bel Air the urban 31 - 37 Patte d'Oie area Pikine Ouest 38 - 44 Thiaroye Diaksao (overcrowdi Malika Guinaw Rail Nord 45 - 50 Guinaw Rail Sud ng ) 51 - 55 Parcelles Assainies Thiaroye Gare 56 - 61 Sanghalkm Wakhinane Nimzatt Pikine Nord 15*8m plots 62 - 70 Pikine Est Ndiareme Limamoulaye Yeumbeul Nord Keur Massar Diamaguene Sicap Mbao Pikine Ouest Sam Notoire camberene Djidah Thiaroye Kaw camberene Medina Gounass Medina Gounass Golf Sud Yeumbeul Sud Djidah Thiaroye Kaw Golf Sud Low septic Sam Notaire tank/pit Parcelles Assainies pikine Nord Wakhinane Nimzatt Pikine Est Thiaroye Gare Ndiareme Limamoulaye Pikine Ouest density in Guinaw Rail Nord Yeumbeul Sud Diamaguene Sicap Mbao Patte d'Oie Thiaroye Diaksao Yeumbeul Nord Mbao the peri- Guinaw Rail Sud Rufisque Ouest Mbao urban area Dalifort Thiaroye Sur Mer Rufisque Ouest Sanghalkm Keur Massar Malika Hann-Bel Air 3 0 3 Rufisque Nord Km AFRIWATSAN Í -5 15 35 55

Fig 2: Density of septic tank or pit (DIAW, 2017) Fig 3 Avearage density of septic tank or pit

Fig 4: Density of septic tank or pit per district community (DIAW, 2017)

OBJECTIVE II : ASSESSING THE CONTAMINATION PROCESSES

Multi-criteria model of vulnerability of acquifer to fecal contamination

Density of septic tank/ pit latrine per 10000 m2 01 - 10 Criteria 11 - 16 Permeabiliy (OMS,1972 ; 17 - 23 24 - 30 Seck,1989) 31 - 37 38 - 44 Malika - Density of septic tank/pit (L1) 45 - 50 51 - 55 - Groundwater level (L2) 56 - 61 Sanghalkm Wakhinane Nimzatt 62 - 70 Ndiareme Limamoulaye Yeumbeul Nord Keur Massar - Living standards (L3) Sam Notoire Not available now camberene Medina Gounass Golf Sud Yeumbeul Sud - Population density (L4) Djidah Thiaroye Kaw Parcelles Assainies pikine Nord Pikine Est Thiaroye Gare Pikine Ouest - Distribution of wetlands (L5) Guinaw Rail Nord Diamaguene Sicap Mbao Patte d'Oie Thiaroye Diaksao Mbao Guinaw Rail Sud - Housing typology (L6) Thiaroye Sur Mer Dalifort Rufisque Ouest

Hann-Bel Air - Permeability (L7) 3 0 3 Rufisque Nord Km AFRIWATSAN Í

Density of septic tank/pit (L1) Depth of the water (L2) Permeability (L7)

Population density (People/Sq.km) Socio-economic profiles 00143 - 02286 No cover  Peer comparison techniques (Saaty,1977) 02287 - 05140 Very poor class 05141 - 11733 11734 - 15415 Poor class Malika 15416 - 20481 Medium class 20482 - 24294 Malika 24295 - 30256 Superior middle class Wakhinane Nimzatt Sanghalkm Ndiareme Limamoulaye Yeumbeul Nord Keur Massar 30257 - 37597 Upper class Golf Sud Sam Notoire Medina Gounass 37598 - 48457 Wakhinane Nimzatt Sanghalkm camberene Yeumbeul Sud 48458 - 68356 Ndiareme Limamoulaye Yeumbeul Nord Keur Massar Djidah Thiaroye Kaw Parcelles Assainies pikine Nord Pikine EstThiaroye Gare Sam Notoire Pikine Ouest Guinaw Rail NordDiamaguene Sicap Mbao Medina Gounass camberene Ngor Yoff Patte d'Oie Guinaw Rail SudThiaroye Diaksao Mbao Golf Sud Yeumbeul Sud Thiaroye Sur Mer Djidah Thiaroye Kaw Dalifort Grand Yoff Parcelles Assainies pikine Nord Thiaroye Gare Pikine Ouest Pikine Est Sicap Liberté Guinaw Rail Nord Ouakam Dieupeul-Derkle Diamaguene Sicap Mbao Patte d'Oie Thiaroye Diaksao Guinaw Rail Sud Mbao Mermoz-Sacré Coeur HLM Biscuiterie Dalifort Thiaroye Sur Mer Grand Dakar Hann-Bel Air

Hann-Bel Air Rufisque Ouest Fann-Point E- Amitié Gueule Tapée, Fass-Colobane

Medina Rufisque Nord Plateau 3 0 3 Rufisque Est 3 0 3 Km AFRIWATSAN Í Km AFRIWATSAN Í Population density (L4) Living standard (L3)  Each creteria (rating scale)

Wetlands Housing type Dense center Classes Level of risk Scale Pericentre very equipped with housing under equipped

Pericentre equipped, predominance of house in floor 01 Very low vulnerability 1

Suburbs is close with small households Malika Suburban suburbs equipped with large households 02 Low vulnerability 2 Malika Suburbs are far under equipped with large households Sanghalkm Wakhinane Nimzatt Periurban under equipped with, owner predominance Ndiareme Limamoulaye Yeumbeul Nord Keur Massar 03 Medium vulnerability 3 Sam Notoire Sanghalkm Medina Gounass Wakhinane Nimzatt camberene Golf Sud Djidah Thiaroye Kaw Yeumbeul Sud Ndiareme Limamoulaye Yeumbeul Nord Keur Massar 04 High vulnerability 4 Parcelles Assainies pikine Nord Thiaroye Gare Pikine Ouest Pikine Est Sam Notoire Guinaw Rail Nord Thiaroye DiaksaoDiamaguene Sicap Mbao Medina Gounass Yoff Patte d'Oie Guinaw Rail Sud Mbao 05 Very high vlnerability 5 camberene Ngor Golf Sud Yeumbeul Sud Dalifort Thiaroye Sur Mer Djidah Thiaroye Kaw Grand Yoff

Parcelles Assainies pikine Nord Pikine Est Thiaroye Gare Sicap Liberté Pikine Ouest Ouakam Dieupeul-Derkle Guinaw Rail Nord Thiaroye DiaksaoDiamaguene Sicap Mbao Patte d'Oie HLM Guinaw Rail Sud Mbao Mermoz-Sacré Coeur Biscuiterie  Grand Dakar RESULTS Dalifort Thiaroye Sur Mer Hann-Bel Air

Grand Yoff Fann-Point E- Amitié Hann-Bel Air Gueule Tapée, Fass-Colobane - Decisional maps and graphs of vulnérability of acquifer to Dieupeul-Derkle Medina

Plateau HLM 3 0 3 3 0 3 AFRIWATSAN Km AFRIWATSAN Í August 2016 Km Í faecal contamination 9 Wetlands (L5) Housing type (L6) - History of the pollution state / Vulnerable Zones OBJECTIVE II : ASSESSING THE CONTAMINATION PROCESSES

Statistical model (Association between NO3 Conc and density of Septic tank / pit )  Extract from the database

Density of ST Grid number Nitrate concentration (mg/L) 1 1 20 5 2 108 3 3 85 10 4 128 1 5 108 - X2 33 6 218  Analysis chi-square test  Linear regression model - P-value

I = b – c*d, I : Nitrate concentration : dependent variable, d: Density of septic tank : independent va…  Many Scenarios How density of septic I = 0 d = b/c tank is appropriate before implanting a I < 50 mg/ L supply source ? Fig 1: Relationship between nitrqte concentration and density of septic tqnk

10 Fig 2: Relationship between nitrqte concentration and density of septic tqnk OBJECTIVE II : ASSESSING THE CONTAMINATION PROCESSES

Statistical model (Association between PO4 Contamination and Septic tank density/Pit)  Extract from the database Phosphate Con (mg/L) Density of septic tank 0.39 7 0.38 3 0.68 13 0.90 25 1.14 40 1.30 42 X2 n n  Analysed with chi-square  Linear regression model P-value

I = b – c*d I : Phosphate concentration : dependent variable, d: Density of septic tank  Scenarios I = 0 d = b/c. How density of septic tank is apprprpriate before I < 5 mg/ L implanting a supply source ? Fig 3: Relationship between phosphate concentration and density of septic tqnk/pit

11 Fig 4: Relationship between phosphate concentration and density of septic tank/pit OBJECTIVE I : HYDRODYNAMISM AND SANITATION FAILURE Uptodating a regional model with GMS - Water-Table Fluctuation models -

- Contribution of

Pambal injection of septic tank (Pumping) and salt intrusion

SEBIKOTANE Boundaries - Boundary conditions Imposed potential h = (δf / δf)

DAKAR Non-flux boundary (Neuman) - an imposed potential h Marly substratum

Tanma Lake - a no-flow (Neumann) Quaternary Sands Aquifer

4 2 0 4 Clay and volcanic rock - an imposed flow r Km AFRIWATSAN Í Fig 5: Correlation of simulated and measured heads (Forward Model)

 (2/3) Recharge method estimation 1. Water-Table Fluctuation models 2. Injection rate of septic tank/pit 3. Rainfall recharge by isotopic method (O16, H2)

1. Water-Table Fluctuation models (Curthbert & al., 2010)

Piezometry (1948) and Tharoye bassin - Distance from observation well to downstream perennial stream: 921.11 m Fig 6 : Annual Determination Coefficient (Inverse Model) - Characteristic distance from groundwater divide to perennial stream: 3812.99 m - Characteristic distance from lateral groundwater divide to ephemeral stream: 13888.4 m Results Input - Specific yield: 15.5 % R = 0.00001187P2 + 0.17374P + 0.01629. - Transmissivity: 24.3 m2/day - Potential evapotranspiration coefficient of determination of 0.905. In other words, - Precipitation Recharge is 17.3% precipitation. 12

OBJECTIVE I : HYDRODYNAMISM AND SANITATION FAILURE

Etablish a regional model with GMS Injection rate of septic tank / pit estimation

2. Injection rate of septic tank/pit High injection rate in - the number of people per house on average the urban area (Max : [ANSD,2016;DIAW,2017] 787 mm/year ) - the consumption of water per people per day is 40 to 68 l / day [OCDE (2006/2007), ASME, 2014, SDE,2014]. - the proportion of water injected into the septic tank which could reach 71 % of the per capita consumption (BELGAQUA, 2008)

- the amount of urine injection which is 1.2 l / day (Mara, 1996) Injection rates ( Peri- urban area) range - the runoff coefficient of 0.8 - 0.95 of cement mortar or cement from 3.4 to 183 mm / concrete of the bottom of septic tank (AFNOR, 2013) year.

Fig 6: Distribution of injection rate of septic tank at the level of Thiaroye acquifer

High injection rate Low injection rate

Fig 6: Distribution of injection rate of septic tank at the level of Thiaroye acquifer

13 Fig 7: Distribution of injection rate of septic tank dépendant on their density at the level of Thiaroye acquifer OBJECTIVE II : ASSESSING THE CONTAMINATION PROCESSES

Establishment of miniature model with GMS

Define a horizontal separation ''Septic tank – supply source'‘ - Semi-analytical particle tracking method (MODPATH) Injection rates of septic tank/pit  The pilot site ≤ 1 or 1.5 km2  Meshing = 10 m  Recharge Construction of shallow piezometer array (3) Rainfall recharge (O18, H2) - A reconnaissance campaign of Hdrogeological team

Pilot area

Pz3 Pz2 Pz1 Borehole 22

- Pumping test, key hydrogeological - Water level parameters and and boundaries conditions - Tracer tests 14 OBJECTIVE III : Monitoring of fecal contamination indicators (NO3, PO4, E.coli) and Assessing the impact of water quality on human health - P2.6 localisation ( Density of - Increase in level (9 21 h) High injection rate (the day) Very shallow character septic tank : 46 /10000 m2) - Decrease in level (21 9 h) Low injection rate (night) - Regular daily evolution

 Monitoring - DW109, Piezometer (Low - Collection of PO4, NO3 density of septic tank and E.coli samples (Every - New Piezometer (High 30 days) density) Chemical evolution

- Next DW109 (Low - Measurement of WTF during dry density ) season for monitoring man induced - P2.6 (High density) recharge (Dataloggers)

Man-induced recharge processes

- Assessing the impact of water quality on human health 15 Conclusion

 Vulnerability of the aquifer to faecal contamination in order to propose different types (urine diversion, ecological latrines with compost, latrines on a mound, latrines perched) according to the degree of Vulnerability

 Formulation of a "Safe Distance" which is a technical prescription crucial for any planner of ecological on-site sanitation systems

 The density of septic tanks or pit latrine adequate not to exceed the WHO standards recommended for drinking water (Facilitate strategic orientations for the relevant government institution)

 The impact of groundwater level and marine intrusion in the horizon (t) according to this scenarios : - Pumping at 100 % on-site sanitation coverage rate and Pumping with a diminution of coverage rate of on-site sanitation.

All outcomes will serve not only as a model for other cities of the Senegal, which the impact of on- site sanitation is in the embryonic stage, but also for major cities of ECOWAS, sub-Saharan Africa and all over Africa. 16 THANK YOU FOR

LISTENING 17