Assessing uncertainty in material flow analysis to support sanitation planning in the global south – the case of Maputo, Mozambique

Material flow Analysis (MFA) can serve as a decision support tool for sanitation planning, since it offers a comprehensive and analytical account of water and nutrients within a defined region. MFA is very data-intensive and... [ view full abstract ]

Material flow Analysis (MFA) can serve as a decision support tool for sanitation planning, since it offers a comprehensive and analytical account of water and nutrients within a defined region. MFA is very data-intensive and data scarcity is a known issue associated with the approach. This shortcoming, especially of relevance in the context of the Global South must be taken into consideration. Otherwise, calculations can lead to over- or underestimation of the modelled variables. Therefore, it is essential to assess the robustness of the produced MFA results in order to be able to take more deliberated and safe decisions. In a mathematical MFA (MMFA) study, the uncertainty of the parameters and variables can be evaluated through Monte Carlo simulation. Furthermore, the critical parameters which have high impact on variable flows can be determined with the help of a sensitivity analysis.

This study demonstrates how MMFA can be applied to the case-study of Maputo, the capital of Mozambique. The model simulates water and nitrogen flows in the area for the year 2015. For the calculations SIMBOX was used, this program is especially fit to problems with limited data availability and uncertainty, by defining the value of input parameters as probability distributions instead of point values. SIMBOX also displays the ranking of parameter knowledge, which consists on a list of ranked input parameters with respect to their impact on the uncertainty of the variable calculated. The developed model is composed of 158 parameters, with their respective attributes, averages, standard deviations and distributions. Most input data were extracted from the literature and from interviews with local water and sanitation experts.

The results show that approximately 4,500 ton of nitrogen reach the groundwater in 2015 and OSS (On-Site Sanitation) infiltration is the main source of this pollution.  The sensitivity analysis applied for the variable “ nitrogen reaching GW (Ground Water)” shows that four parameters - Fraction of N in food protein, Total food protein, Fraction of septic tanks legally discharged and City’s population - explain almost 93% of the uncertainty associated. As the City’s population in 2015 is certain, further efforts need be put to the knowledge on ‘Total food protein’, ‘Fraction of  N in food protein’ and ‘Fraction of septic tanks legally discharged.  A simulation after improving the uncertainty of the parameters to a maximum 10% standard deviation showed now that four parameters explains 84.4 % of the uncertainty for the variable “ nitrogen reaching GW (Ground Water)”  , dropping 8.5% from the previous ranking of parameter knowledge.