Review Papers Claim by Claim

What the authors did (short)

• Computed household consumer water footprints (green+blue+grey) using bottom‑up household survey food quantities (INS) for 1980–2010 and WFN product WFP values, aggregated to per‑capita WFP by product group and region Souissi et al. 2019 methodology
• Key reported empirical numbers: total consumer WFP ~1,208 m3/cap/yr in 1985 → ~1,586 m3/cap/yr in 2010; cereals contribution fell while meat and dairy rose substantially; urban WFP > rural WFP; ~71% internal vs ~29% external average over study period Souissi et al. 2019 results

Critical appraisal — strengths

• Bottom‑up, household‑level approach: adds behavioral and regional resolution missing from national balance‑sheet/ top‑down studies and allows urban/rural and region comparisons Souissi et al. 2019 novelty
• Use of WFN spatially explicit green/blue/grey breakdowns is standard in the field and enables policy‑relevant internal vs external (virtual water) analyses.

Critical appraisal — key limitations & blindspots

• Reproducibility gap: the paper does not publish the full per‑household calculation tables, WFN product lookup tables used, nor the exact mapping (INS product codes → WFN product IDs). That prevents independent re‑runs and sensitivity checks of core numbers (missing raw calculation tables) Souissi et al. 2019 data availability
• Sensitivity to WFN parameter choices: WFN WFPs embed production‑system averages (yields, irrigation shares, leaching factors). Small changes in assumed yields or irrigation fractions can move blue/green shares and product WFPs substantially — authors cite WFN but do not present sensitivity analysis (a standard robustness check is absent) Mekonnen & Hoekstra 2011 WFP framework
• Temporal mismatch & aggregation: INS survey frequency (every 5 years) and use of week‑weighed consumption extrapolated to annual values can introduce sampling or seasonal bias (surveys may under/over sample certain seasons); authors acknowledge survey limitations but do not quantify seasonal bias.
• Nutrition vs WFP tradeoffs under‑explored: paper recommends shifting diets (less beef) but does not quantify nutritional impacts or affordability tradeoffs of proposed substitutions — an important blindspot for food security policy recommendations.

How much do conclusions depend on assumptions?

Major dependence points: WFN product WFPs (yields, irrigation shares), treatment of feed imports (how much of animal WFP is allocated to feed vs animal product), and mapping of INS food codes to WFN categories. Any change in these can shift internal vs external shares and the meat vs cereal contributions materially.

Recommendations to strengthen paper & follow‑ups

1. Publish the full calculation matrix (INS food quantities × WFN per‑product WFP) as supplementary data or repository (enables reproducibility and third‑party audits).
2. Run and report sensitivity analyses: vary crop yields, irrigation fraction, grey‑water leaching fraction and feed import ratios to produce uncertainty bounds on per‑capita WFP and internal/external shares.
3. Integrate nutritional/affordability modelling: when recommending meat reduction, quantify caloric/protein tradeoffs and realistic substitution pathways (poultry/fish/legumes), and estimate policy impacts on the poorest deciles.
4. Consider scenario modelling (2030/2050) coupling diet/urbanization projections with climate change impacts on yields to evaluate future WFP trajectories under alternative policies.