**1. Introduction**

Sustainable development, as defined by the United Nations, is "development that meets the needs of the present without compromising the ability of future generations to meet their own needs" [1]. As an essential primary natural resource, the role of water in sustainable development is well recognized due to its critical functions in social and economic activities [2]. Unsustainable use of freshwater resources by nations can hamper their food security and long-term economic development [3]. The future water vision emphasizes that by 2050, "global markets and trade flows would be monitored through a global water sensitivity certification scheme that ensures water-intensive products are exported from areas with comparatively little or no water stress. We will recognize the economic value of water and all forms of an economic enterprise will take consideration of the water implications of their actions" [2]. Virtual water trade (VWT) can be used to gauge the water that is virtually transferred from one region to another region through commodities and services, both at global and national levels.

Nations can ensure the sustainable use of domestic freshwater by prioritizing their production and trade decisions based on water availability. Trade in virtual water is the amount of hidden/virtual water that crosses borders with commodities/services. Virtual water content (VWC), defined as the ratio of the total water used for the production of the crop to the total volume of crop produced (m3/ton), is a crucial concept used in the relevant literature [4]. The virtual water content of primary crops is calculated based on crop water requirements and yields [4]. Several studies have demonstrated the role of virtual water trade in ameliorating regional water shortages. Chapagain et al. [5,6] found that virtual water trade saved 6% of global agricultural water use. More recently, References [7–10] demonstrated the role of virtual water in improving water use efficiency and saving global water resources. Several national-level studies have also shown that water-scarce countries have used VWT to alleviate pressure on domestic water resources by importing water-intensive products [11–18]. In his seminal studies on virtual water, Allan [19,20] discussed how the Middle Eastern countries were able to overcome national water shortages through cereal imports. Many other important studies on Brazil, China, Cyprus, Egypt, the Middle East, Morocco, Spain and Tunisia also show the role of VWT in saving (and losses, if any) of domestic water resources for these countries [11–14,21–29].

Pakistan—with the world's fifth-largest population [30]—is a water-stressed country and its freshwater resources face heightened pressure from increasing population and climate change [31]. With an average annual rainfall of about 250 mm, Pakistan is among the most arid countries of the world [32]. Relying upon the most extensive contiguous irrigation system in the world, Pakistan uses over 94% of national water withdrawal in agriculture [33]. In the past, the country failed to add any large water reservoirs, thus adding to the pressure on dwindling water resources, especially in agriculture [34]. Due to declining per capita availability of water, Pakistan will turn from currently water-stressed to a water-scarce country in the 2040s [31]. Groundwater depletion due to unsustainable use of the aquifers for agriculture is becoming more severe in Pakistan [35]. Moreover, increasing quantities of pesticides and fertilizers used in agriculture are causing large-scale uncontrolled pollution of surface and groundwater [31].

Agriculture, the most water-intensive sector in Pakistan, can be taken as the strongest candidate to overcome the growing national water scarcity. Over 26% of crop-related groundwater depletion in Pakistan is due to crops exported to other countries, the majority of which (82%) is embedded in rice exports [35]. Although several studies have discussed the relevance of virtual water in agricultural trade for Pakistan [6,35–38], most of these studies either fail to distinguish between blue and green water (the blue water includes surface water and groundwater, while the green water is the rainwater stored in the soil as soil moisture [37]) or are related to domestic production/consumption of blue/green water. Chapagain et al. [6] report net savings of Pakistan's water resources through the international trade of all major crops and livestock products. However, the study is limited with respect to reporting only point analysis (average for 1997–2001) and reporting only total water savings. Dalin et al. [35] show that Pakistan is facing groundwater depletion through the international food trade. Although an influential study, it is limited to groundwater and presents its analysis only for two years, that is 2000 and 2010. The study by Fraiture et al. [36] shows net water savings for Pakistan. The study only covers total water savings through the cereal trade in 1995. Chapagain and Hoekstra [37] report blue, green and greywater impacts of international trade of rice for Pakistan and other countries for the period 2000–2004. This study, however, lacks a temporal analysis and the coverage of other agricultural commodities. The green, blue and grey water footprint of global crop production, including that of Pakistan, are covered in Reference [38]; however, no trade analysis is presented in the study. Pakistan's high dependence on freshwater for agriculture production, dwindling water resources and increasing trade in agriculture should be a strong motivation for conducting a more detailed study on trade in virtual water for Pakistan.

The primary purpose of this study is to assess the evolution of trade and savings/losses of Pakistan's blue and green virtual water through international trade of the agricultural commodities during 1990–2016 and in 2030. By doing so, we can answer the following research questions: What are the major commodities and source/destination regions involved in Pakistan's VWT? How Pakistan's VWT has evolved over the years and which policies have driven the changes in VWT? More importantly, was the trade dominated by blue or green virtual water? Has Pakistan been saving blue/green water

through its trade in agricultural commodities at the national and global levels? Has Pakistan's trade in blue and green VW been economically viable over the years? How would the trends evolve in the future? What is the value of the virtual water contents of these agricultural commodities? A detailed analysis of historical and future blue and green virtual water movements through Pakistan's global agricultural trade can help policy planning on water use for the country. To the best of our knowledge, this is also the first study focused on a water-stressed country that exports large quantities of agricultural products and virtual water.

The rest of the paper is organized as follows. Section 2 describes the methods and data used in this study. Section 3 presents our results on virtual water contents of these commodities for Pakistan; and trade and net savings/losses in terms of total, blue and green virtual water over the past years and in the future. Section 4 discusses the implications of our results and presents policy and research recommendations. The Section 5 presents the conclusions.
