Electronic Supplementary Material

Note: Section numbers refer to the main paper’s sections to which the information is being added.

1-Introduction

Fig.S1: General water use impacts framework (adapted from Kounina et al). The dotted rectangle highlights the impact pathway for which methods are compared in this paper.

Table S1: Summary of methods and their names

Impact assessed / Reference / Name / Details
Scarcity / Frischknecht, 2008 / M-SwissSc / Midpoint, scarcity, withdrawal-to-availability, power function
Pfister, 2009 / M- PfisterSc / Midpoint, scarcity, withdrawal-to-availability, logistic function
Hoekstra, 2012 / M-BWSSc / Midpoint, scarcity, consumption-to-availability, direct function
Boulay, 2011 / M-BoulaySc / Midpoint, scarcity, consumption-to-availability, logistic function
Availability / Boulay, 2011 / M-BoulayAv / Midpoint, availability, consumption-to-availability (quality specific), logistic function
Veolia, 2010 / M-WIIXAv / Midpoint, availability, withdrawal-to-availability and distance to target for pollution
Human Health / Pfister, 2009 / E-Pfister / As published, agricultural deprivation
Motoshita, 2010a / E-Motoshita_dom / Effect factor as published, domestic deprivation, then combined with M-PfisterSc and distribution factor (DAU)
Motoshita, 2010b / E-Motoshita_agri / Adapted from presentation, agricultural deprivation including trade effect , then combined with M-PfisterSc and distribution factor (DAU)
E-Motoshita_agri (no TE) / Adapted from presentation, agricultural deprivation excluding trade effect , then combined with M-PfisterSc and distribution factor (DAU)
Boulay, 2011 / E-Boulay_marg / Simplified from publication (no quality), considers agriculture as off-stream user deprived (100%) and aquaculture as in-stream user deprived.
E-Boulay_distri / Simplified from publication (no quality), considers off-stream users to be deprived proportionally to their use (agriculture and domestic are included) and aquaculture as in-stream user deprived.
E-Boulay_agri / Simplified from publication (no quality), represents a partial factor from E-Boulay_distri for comparison purposes, only for agricultural users.
E-Boulay_dom / Simplified from publication (no quality), represents a partial factor from E-Boulay_distri for comparison purposes, only for domestic users.
E-Boulay_marg_Q / As published, considers agriculture as off-stream user deprived (100%) and aquaculture as in-stream user deprived.
E-Boulay_distri_Q / As published, considers off-stream users to be deprived proportionally to their use (agriculture and domestic are included) and aquaculture as in-stream user deprived.
E-Boulay_agri_Q / Represents a partial factor from E-Boulay_distri for comparison purposes, only for agricultural users.
E-Boulay_dom_Q / Represents a partial factor from E-Boulay_distri for comparison purposes, only for quality users.

3.1.1Scarcityindicators

Fig.S2 plots the normalized scarcity indicators from all methods against the WTA index. The underlying data used to calculate this latter (WaterGap), differs from the ones used in M-SwissSc, which explains the observed inconsistency of the data series. Please note that all values equal zero (60% of values for M-Boulay-Sc), cannot be shown on the graph. The average values described in section 2.3 is also shown.

Fig.S2: Log graph of normalized scarcity methods against WTA for the main watersheds showing spread and differences among methods.

3.1.4Human health: Domestic user deprivation

Fig.S3: Comparison of human health model outcomes from domestic water deprivation impact pathways using Boulay and Motoshita models. A) CF and b) socio-economic and effect factor (SEE), which excludes scarcity and distribution of affected users.

3.1.5Human health: Agricultural user deprivation

While the consistency is higher between E-Motoshita_agri without trade effect and the two other models then it is with trade effect included, the mean difference between models decreases in comparison with E-Pfister but increases when comparing with E-Boulay_agri. This is because E-Boulay_agri shows higher results in general, and the addition of the trade effect increases the results as well in E-Motoshita_agri.

Fig.S4: Comparison of agriculture water deprivation impacts on human health. a) complete CF, b) Socio-economic and effect factors and c) effect factors only.

3.1.6Inventory-related choices

Temporal variations

Fig.S5: Difference between annual scarcity indicators calculated from annual data vs. from a withdrawal-based weighted average of monthly data. Results are obtained with M-PfisterSc which scarcity indexes range from 0.01 to .1

Water Source

Fig.S6: Absolute difference in scarcity results between general scarcity indicators calculated using all water use and availability and weighted-average of surface and groundwater scarcities, based on intensity of groundwater withdrawals. Results are obtained using M-Boulay-Sc, which values range between 0 – 1.

Quality aspect

Out of the 600 regions of the world for which data was available (based on Boulay et al[30]), scarcity indicators are higher when consuming water of good quality then of unspecified quality in 42% of the cases. No difference is observed for the rest of the cases (58%), representing regions where general scarcity is already maximal (value of 1). Consuming water of poor quality results in higher stress in about 21% of the cases, corresponding to region where water quality is very poor. No difference is observed in 79% of cases, i.e. in regions where the average water quality is poor. Degrading good quality into very poor quality,will result in higher impacts than consuming general water (quality non-specified) in 39% of the cases, in countries with no general water scarcity, but with good quality water scarcity. These will present impacts when considering water of a certain quality but not when considering all water available.

3.1.7Scarcity

Fig.S7: Different model functions defining scarcity as a function of CTA a) normal scale and b)log scale.

3.2.3

Fig.S8: Difference of including or not domestic users

1