**5. Conclusions**

The quantification of virtual water trade provides an essential perspective on sustainable water use practices. Using Pakistan as a case study, this paper assessed the trade and savings/losses of blue and green virtual water through agricultural commodities, over the period 1990–2016 and in 2030. The results of our study show that in most of the studied commodities, blue VW is the major component in the total water use. Pakistan's export of VW increased more rapidly than the import of VW over 1990–2016. On average, the net VW import (import minus export) was positive but small. Pakistan has been a net exporter of blue VW, mostly through rice export to Asian and African countries. In terms of green VW, Pakistan has been a net importer, mainly through the import of palm oil from Indonesia and Malaysia.

In terms of domestic savings, Pakistan's trade in the agricultural commodities has been saving green VW for the country. However, Pakistan has been losing increasingly higher volumes of blue VW through its agricultural trade over the study period. Putting this into perspective, the blue VW loss was equivalent to about 7% of the total agricultural water withdrawal in the country in 2011. Pakistan's trade in the agricultural commodities has also contributed to positive savings of blue and green virtual water at the global level.

In the future (2030), both Pakistan's domestic savings of green and losses of blue virtual water will increase by more than 200%. The global savings will increase by more than three times for green VW and more than 120% for blue VW. Our results also suggest that Pakistan has been exporting more expensive (with high opportunity cost) blue VW through its agricultural trade to the rest of the world and can thus benefit from improvements in water use efficiency (in the export-oriented crops) and adjustment in its export portfolio of agricultural commodities by promoting the export of commodities with higher value and lower water use intensity.

As discussed in detail, two major shortcomings should be addressed in future research. First, time-series data from the national model(s) should be used in place of the global estimates for VWC. This will enhance the reliability of the results to produce more reliable policy suggestions. Second, time-series data for grey VW for the agricultural commodities are needed to analyze the environmental effects of trade in agricultural commodities, which were not taken into account in this study.

Although numerous studies on VW trade have been published, the studies mainly concerned those countries which are either net-importers and water-stressed or net-exporters and water-abundant. Given that water availability is more crucial for the water-stressed net-exporting countries, the trade in VW for such countries becomes more critical. The analysis of Pakistan's trade in VW through agricultural commodities presented in this study can further promote future work in water-stressed countries.

**Author Contributions:** Conceptualization, T.A. and W.X.; Formal analysis, T.A. and W.X.; Methodology, T.A., A.M.N. and M.F.R.; Writing—original draft, T.A. and W.X.; Writing—review & editing, T.A., A.M.N. and M.F.R. All authors read and approved the final manuscript.

**Funding:** The authors acknowledge their respective financial supports from the National Key R&D Program of China (2016YFA0602604) and the National Natural Sciences Foundation of China (71503243; 71873009).

**Conflicts of Interest:** The authors declare no conflict of interest.
