The relationship between on-site sanitation density and shallow groundwater quality: evidence from remote sensing and in situ observations in ,

Mor Talla Diaw, PhD Student Université Cheikh Anta Diop, Senegal www.afriwatsan.org

funded with UK aid from the UK Government through the Royal Society – DFID Africa Capacity Building Initiative BACKGROUND TO THE RESEARCH Water management in Senegal is experiencing multiple issues : Degradation of water quality by on-site sanitation systems failure… JEUDI 12 AVRIL 2018

on-site sanitation (pit latrines, septic sewered sanitation systems serve a tanks) most widely system in low- minority of urban dwellers in most low- income cities of Africa income cities of Africa

Planning or readjustment of improved supply sources & low cost on-site sanitation

achieving UN SDG 6 of universal access to safe water and sanitation by 2030 in African cities relies upon on-site solutions OBJECTIVE AND METHODOLOGY OF OUR STUDY

 OBJECTIVE Evaluate the density of on-site sanitation facilities associated with faecal contamination of groundwater below WHO guidelines for safe water

 METHODOLOGY

. City-scale mapping of on-site sanitation fafilities (i.e., septic tanks, pit latrines)

. Sampling of shallow groundwater sources for faecal contaminants

. Statistical modeling (Chi-squared tests, Spearman correlation ….)

. Validation of statistical models

. Implied outcomes 3 SHALLOW AQUIFER AND ON-SITE SANITATION AREA

Built Built area: Malika

Sanghalkm Wakhinane Nimzatt Ndiareme Limamoulaye Yeumbeul Nord Keur Massar ~80% of the

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

Rufisque Nord area Sanghalkm Sicap Liberté Malika Dieupeul-Derkle

Wakhinane Nimzatt Mermoz-Sacré Coeur HLM Ndiareme Limam oulayeYeumbeul N ord Keur Massar Sam Notoire Hann-Bel Air camberene Medina Gounass Golf Sud Yeumbeul Sud Djidah Thiaroye Kaw pikine Nord Parcelles Assainies Thiaroye Gare Fann-Point E- Amitié Pikine Est Gueule Tapée, Fass-Colobane Pikine Ouest Diamaguene Sicap M bao Mbao Yoff Patte d'Oie Ngor Thiaroye Sur Mer Medina Dalifort Rufisque Ouest Grand Yoff

Rufisque Nord Sicap Liberté Ouakam Dieupeul-Derkle Plateau 3 BiscuiterieHann-Bel Air Mermoz-Sacré Coeur HLM Grand Dakar 3 1.5 0 Km AFRIWATSAN Fann-Point E- Amitié Í Gueule Tapée, Fass-Colobane Medina Thiaroye shallow aquifer r Plateau Sewered sanitation area

7 3.5 0 On-site sanitation area Study area AFRIWATSAN Km Í - Urban aquifer provided 10% of the water supply until excessive nitrate concentrations – traced to faecal sources using 15N/14N shallow aquifer of Quaternary sands (Diediou, 2012), restricted further use of urban wells in 2008 on peninsular Senegal

On-site sanitation Sewered sanitation On-site

Cité Keur Salam " sanitation Malika systems

Wakhinane Nimzatt Sanghalkm Ndiareme Limamoulaye Yeumbeul Nord Keur Massar (septic tanks Sam Notoire Medina Gounass camberene Golf Sud Yeumbeul Sud Djidah Thiaroye Kaw

Parcelles Assainies pikine Nord Thiaroye Gare or pit 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 latrines): Grand Yoff

Rufisque Nord Sicap Liberté Ouakam Dieupeul-Derkle coverage

Mermoz-Sacré Coeur BiscuiterieHLM

Grand Dakar Hann-Bel Air rate

Fann-Point E- Amitié Gueule Tapée, Fass-Colobane (90 - 95%) Medina

Plateau 3 0 3 Km AFRIWATSAN Í COLLECTION AND ANALYSIS OF FAECAL CONTAMINATION INDICATORS A network of 48 sampling wells consisting of dug-wells, piezometers and boreholes (37) 1st 1st campaign !. Piezometers Area affected by (19) 2n 2nd campaign faecal contamination "/ Deug wells (25) 3r 3rd campaign 0# Boreholes

! ! Faecal ! ! ! ! ! ! ! ! !! !! ! ! ! ! ! ! ! ! ! ! !! indicators ! ! ! ! ! ! ! !!

DW01 - DW06 "/ - [NO ] (laboratory) "/ 3 DW07 "/ DW04 DW8 "/ DW26 "/ "/ DW03 - E. coli counts (laboratory) DW06 DW05 DW25 F32 "/ P02.5 DW21 "/ "/ DW24#0"/ !.#0 DW22!. DW14 DW9 F30 "/ F22 "/ DW20"/ "/ DW23#0 #0 F33 "/ DW19 "/ "/ DW18"/ DW02 #0 Analyses #0 F18 F17bis"/ DW11 P2,2 #0 DW12 !. "/ "/ DW10 - "/ - [NO3 ] Cd reduction and P03.1 photometry !. DW17 DW15 DW13 "/ "/ "/

2.5 1.25 0 Km AFRIWATSAN Í - [- E. Coli counts FISH5

RESULTS FROM SPOT SAMPLING OF GROUNDWATER

53 % 70 % of of wells > Sampl 0 E.coli ing wells /100ml > 50 ; mg/L

Max[E. coli count ] Max[ = 654 E. NO3-] coli/10 = 710 0 ml mg/L

Post-monsoon sampling (Aug 2016) shows both increased contamination (Injection rate of septic tanks or pit latrines ) and reduced contamination (dilution from recharge rain-fed) relative pre-monsoon (Apr 2017)

800

700 August 2016 April 2017

600

500

] - 3 400

[NO 300

200

100

0 P2-5 P2-6 P3.1 HT2 HT5 HT6 HT7 DW4 DW1 DW10 DW12 DW14 DW16 DW2 DW3 DW5 DW58 DW58bis DW8 6DW9 MAPPING OF ON-SITE SANITATION FACILITIES 1

 DISTRIBUTION OF SANITATION TYPE / HIGH RESOLUTION SATELLITE IMAGE USED

2 On-site sanitation Source: National institution - western zone, 74 km for sanitation (ONAS) Sewered sanitation (sewered sanitation)

Cité Keur Salam " - eastern zone, 99 km2 Malika (on-site sanitation)

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 Thiaroye DiaksaoDiamaguene Sicap Mbao  Mbao High resolution Yoff Patte d'Oie Guinaw Rail Sud Ngor Dalifort Thiaroye Sur Mer Rufisque Ouest Satellite image used Grand Yoff

Rufisque Nord Sicap Liberté Data type Acquisition Source Resolution Ouakam Dieupeul-Derkle date (m)

Mermoz-Sacré Coeur BiscuiterieHLM Satellite December 2016 QuickBird 1 Grand Dakar Hann-Bel Air image

Fann-Point E- Amitié Gueule Tapée, Fass-Colobane Satellite 2009 Orbview 1.60 image Medina

Plateau HR satellite 2012 Geoeye 0.60 3 0 3 Km image AFRIWATSAN Í MAPPING OF ON-SITE SANITATION FACILITIES 2

 Meshing / Object-oriented classification of high resolution satellite image

Meshing of 100/100 m at the scale unsaturated zone of the shallow aquifer

 Object-oriented classification (Spectral values and morphological parameters)

 A rough estimation of surfaces occupied by the houses on 100/100 m MAPPING OF ON-SITE SANITATION FACILITIES 3

 REFINED AND VALIDATED BY PHOTO-INTERPRETTION AND TRUTHING SURVEYS

DISTRIBUTION OF VALIDATION MESHES VISUALIZATION OF A VALIDATION MESH OF 200/300 m

Center of the Size of the Number of tanks or pit Number of tanks or pit mesh (well X (m) Y (m) validation latrines obtained in latrines obtained in Validation identifier) mesh 200/300 m mesh via 200/300 m mesh through mesh satellite imagery truthing surveys Mesh 01 P.2.2 238560 1632570 200/300 031 027 Mesh 02 F30 242530 1634140 200/300 128 137 Mesh 03 DW18 242790 1633795 200/300 197 205 Mesh 04 F19 244890 1634001 200/300 311 308 Mesh 05 P.3.1 244793 1631188 200/300 093 101 Mesh 06 P.2.6 248000 1633700 200/300 148 154 Mesh 07 DW4 242790 1633795 200/300 116 122 Mesh 08 DW202 257326 1632464 200/300 045 050 Mesh 09 DW109 257203 1638571 200/300 072 078 Mesh 10 DW234 257203 1638571 200/300 083 085 MAPPING OF ON-SITE SANITATION FACILITIES 5  RS analysis strongly correlated to ground-truthing surveys

350 R² = 0,996

300

250

200 obtained spatial imagery obtained spatial imagery in

150 a 200*300 200*300 mesh a

100 of septic septic of tanks pit or latrines

50 Nomber 0 0 40 80 120 160 200 240 280 320 360 Number of tanks or pit latrine obtained via field surveys MAPPING OF ON-SITE SANITATION FACILITIES 4

 Results of advanced on-site sanitation / Density of of tanks or pit latrines

Density of septic tank/ pit latrine per 10000 m2 High septic 01 - 10 tank density 11 - 16 in the urban area 17 - 23

24 - 30 Density max 31 - 37 = 70 tanks 38 - 44 Malika or pit latines 45 - 50 /hectare 51 - 55

56 - 61 SangalkamSanghalkm Wakhinane Nimzatt 62 - 70 Ndiareme Limamoulaye Keur MassarKeur Massar YeumbeulYeumbeul Nord Nord

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

Parcelles Assainies pikine Nord Pikine Est Thiaroye Gare Pikine Ouest Guinaw Rail Nord Diamaguene Sicap Mbao Patte d'Oie Thiaroye Diaksao Mbao Guinaw Rail Sud Mbao Low septic tank density Dalifort Thiaroye Sur Mer Rufisque Ouest in the peri

Hann-Bel Air urban area 3 0 3 Rufisque Nord Km AFRIWATSAN Í RESULTS GENERATED BY STATISTICAL MODELS

 Correlation coefficients of between faecal contaminants and density of tanks/pit

Minimum number Correlation Kendall tau-C ( τ) Gamma Spearman Chi-square (X2) / P- 2 of meshes observed coefficient (R ) / / P-value Coefficient (ϒ)/ P- correlation ( ρ)/ P- value in the study area P-value value value

2 2 R = 0.80 τ = 0.78 ϒ = 0.78 ρ = 0.93 X = 65.550 3,325 meshes 0.000 0.000 0.000 0.000 0.0001

Faecal Nitrate contaminants concentration [NO3 ] : dependant variable)

Linear ρ on-site regression: sanitation = 0.08 * [NO3 ]+ 14 ; ρ: Independant variable 12 REVERSE VALIDATION OF THE STATISTICAL MODELS 2  A very significant association between faecal contaminants and on-site sanitation density

501

R² = 0,80 451 R² = 0,69

401 401

351 latrine per 300/300m latrine per

301

301 pit

251

201 201

151 Density of tanks or tanks of Density

Density of tanks or pit latrines per 300/300m latrinespit per or tanks of Density 101 101

51

1 1 0 100 200 300 400 500 600 700 0 50 100 150 200 250 Nitrate concentration (mg/L) Escherichia coli/100 mL

Use WHO guideline value for nitrate (< 50 mg/L) as a criterion for safe/unsafe water representing the ability of the shallow aquifer to attenuate faecal contamination through dilution and degradation.

Exceeding densities ranging from 18 to 24 on-site sanitation facilities (septic tanks) per hectare in the Thiaroye aquifer pose a serious risk to human health 13 CONCLUSIONS

 density of on-site sanitation facilities has been successfully mapped in a low-income area (99 km2) of Dakar (Senegal) using an object-oriented classification of optical imagery (QuickBird, GeoEye, OrbView) and photo-interpretation, validated by ground- truthing surveys  density of on-site sanitation facilities, comprising primarily septic tanks, is strongly correlated to the magnitude of faecal contamination (nitrate, E. Coli) observed in shallow groundwater within unconsolidated Quaternary sands

 Exceeding densities ranging from 18 to 24 per hectare is associated with nitrate contamination exceeding the WHO guideline value of 50 mg/l

 analysis highlights the capacity of an unconfined shallow sand aquifer to attenuate untreated faecal effluent from on-site sanitation

 replication of these analyses in other conurbations in Sub-Saharan Africa is required to better understand the attenuation capacity of urban aquifer systems in different contexts and to inform use of low-cost sanitation systems to contain faecal effluent 14 Afriwatsan.org Thank you for your listening

Afriwatsan team 15