Water footprint of growing vegetables in selected smallholder irrigation schemes in South Africa

Crop water footprint (WF) is the volume of fresh water used to produce a certain crop in all the steps in the production line. The CROPWAT model was used to calculate crop evapotranspiration, differentiating green and blue water in Zanyokwe (ZIS), Thabina (TIS) and Tugela Ferry (TFIS) Irrigation Schemes. Green beans had the highest water footprint in all three irrigation schemes with 3 535.7 m3/t in TIS, 2 753 m3/t in TFIS and 2 407.6 m3/t in ZIS. Cabbage had the lowest water footprint. The highest water footprint for growing cabbage was 254.5 m3/t in TFIS, followed by 223.1 m3/t in TIS, and the lowest was 217.8 m3/t in ZIS. Green WF represented the highest percentage of water use at ZIS (50.5%), followed by blue water at 26.5% while grey water constituted 22.9%. At TFIS blue, green and grey water use was 23.1%, 56.7% and 20.2%, respectively. The differences observed in the WF of different crops and different schemes were attributed to the differences in weather and environmental characteristics. Green beans had the highest grey water footprint, i.e., 373 m3/t and the lowest was cabbage with 37 m3/t. Potato, spinach and tomatoes had footprints of 156 m3/t, 214 m3/t and 132 m3/t, respectively. For future research it is necessary to consider the possibility and trade-offs of shifting production of each crop to the places where it is most efficient, and to focus on blue water scarcity in each of the case study locations


INTRODUCTION
Fresh water has become a scarce and over-exploited natural resource (UNESCOWWAP, 2006;Ridoutt and Pfister, 2010).Consequently, agriculture is under intense pressure to reduce depletion of water sources, the pollution of water systems and its contribution to soil infertility and erosion (FAO, 2010).According to Fanadzo et al. (2010) and Mnkeni et al. (2010), the decline in available freshwater resources in South Africa is caused and exacerbated by ignorance, and improper measurement and monitoring of water use in smallholder irrigation schemes (SHIS).Therefore, in order to minimise this decline within the South African SHIS, measurements and reports highlighting and addressing high water use in their operations and product life cycles, are necessary (Chapagain and Orr, 2009).
The water footprint (WF) (Hoekstra, 2002;Hoekstra and Chapagain, 2008) is an important tool for considering water conservation impacts from a variety of farm management options (Dourte and Fraisse, 2012).The WF concept expresses the virtual water content of products, organizations, people and nations, in a spatially and temporally explicit way (Hoekstra and Hung, 2002).Dourte and Fraisse (2012) defined WF as the comprehensive measure of freshwater consumption that connects consumptive water use to a certain place, time, and type of water resource.A crop WF is the volume of freshwater used to produce a certain crop in all the steps in the production line.It is a multi-dimensional indicator, showing water consumption volumes by source and polluted volumes by type of pollution; all components of a total water footprint are specified geographically and temporally (Hoekstra et al., 2009).The concept includes three types of virtual water, blue, green and grey.Blue-water footprint (WF blue ) of a crop refers to the volume of irrigation water that evaporates from a crop field during the growing season.Green-water footprint (WF green ) is the volume of rainwater that evaporates from a crop field (Hoekstra, 2009).Grey-water footprint (WF grey ) refers to pollution and is defined as the volume of freshwater that is required to assimilate the load of pollutants based on existing ambient water quality standards (Ercin, 2011).A crop WF is dependent on climatic and specific crop parameters collected over time.
Hitherto, various studies on crop WF have been carried out around the world, where estimation was done for crops such as wheat (Mekonnen and Hoekstra, 2010), cotton (Chapagain et al., 2005), tea and coffee (Chapagain and Hoekstra, 2007) and rice (Chapagain and Hoekstra, 2010).In South Africa, WF research is scant.The few available reports concentrate on domestic consumption, power and processing industries.For example, Unilever (2009) quantified the total water use for a wide range of its products, while SABMiller and WWF (2009) focused on the company's WF in beer production.
On the other hand, studies on SHIS have focused on assessing their performance (Bembridge, 2000, Mnkeni et al., 2010;Crosby et al., 2000).Yokwe (2009) evaluated the performance of the Zanyokwe Irrigation Scheme (ZIS) from an economic point of view, while Shongwe (2007) evaluated the water distribution at the Tugela Ferry Irrigation Scheme (TFIS).Low yield levels and limited knowledge of crop production were identified as one of the constraints to improved crop productivity (Fanadzo et al., 2010).Machete et al. (2004) andMnkeni et al. (2010) identified water mismanagement as the main agronomic factor limiting productivity in ZIS.Mnkeni et al. (2010) further noted little control in the usage of water, resulting in problems of over-application.According to Ntsonto (2005) good irrigation management is a problem in many schemes with ZIS farmers not using water efficiently.A study by Yokwe (2009) highlighted the need to improve water productivity in smallholder irrigation, while Mnkeni et al. (2010) stated that priority should be put on the scope of increasing water productivity at irrigation schemes by ensuring its availability and effective distribution within the schemes.
This research was undertaken to estimate water footprints of vegetable production under smallholder agriculture, this will generate information for further reducing poverty and sustainable development.Particular attention was given to the green-, blue-and grey-water footprint of 5 vegetable crops in the production cycle at ZIS, TFIS and TIS in the Eastern Cape, KwaZulu-Natal and Limpopo provinces, respectively.The estimates were done in a spatiallyexplicit way, from a production perspective, using a conceptual framework based on the FAO CROPWAT approach (Allen et al., 1998)

METHODS AND DATA
This study focused on the production stage, that is, the cultivation of the vegetables, from sowing to harvesting, during the 2000-2013 period.Crop WF estimates were calculated following the Crop Water Requirement (CWR) option, which estimated evapotranspiration under optimal conditions, which means that crop evapotranspiration (ET c ) equals CWR (Hoekstra et al., 2011).WF blue and WF green of vegetable crops produced at each scheme were quantified by taking into account local climatic and soil conditions in the respective provinces.Local nitrogen fertilizer rates applied were used to calculate the grey water.Crop evapotranspiration was calculated using a grid-based dynamic water balance model, which computes a daily water balance and calculates crop water use and requirements.Computations of vegetable crop evapotranspiration and yield were carried out following the method and assumptions provided by Allen et al. (1998).The CROPWAT 8.0 model, a computer program for the calculation of crop water and irrigation requirements based on soil, climate and crop data, was used to estimate crop water use.It calculates reference evapotranspiration, crop water requirements and crop irrigation requirements.

Study area
The climate in the ZIS scheme is semi-arid and relatively mild, with a mean annual rainfall of about 580 mm of which about 445 mm is received in summer.Tugela Ferry Irrigation Scheme is situated in a dry to semi-arid zone, with a mean rainfall of 600-700 mm per annum and very high summer temperatures of up to 44°C (Cousins, 2013).Annual rainfall averages at TIS are 790 mm with drastic inter-annual variations, recurrent droughts and long dry seasons (Jordaan and Grove, 2012).Monthly rainfall and temperature averages for the study period 2000-2013 are shown in Fig. 1.

Evapotranspiration (ET c )
The actual crop evapotranspiration (ET c , mm/day) was calculated using Eq. 1 (Allen et al., 1998, Mekonnen andHoekstra, 2011). (1) where: K c is the crop coefficient, K s [t] a dimensionless transpiration reduction factor dependent on available soil water; ET O [t] the reference evapotranspiration (mm/day).The value of K s was calculated on a daily basis as a function of the maximum and actual available soil moisture in the root zone.The amount of rainfall lost through runoff was computed using Eq. 2, as in the Hydrologiska Bryans Vattenbalansavdelning (HBV) model (Bergstrom, 1995;Lid and Harlin, 2000;Mekonnen and Hoekstra, 2011).
where: RO [t] is the runoff on day t [mm]; P [t] is the precipitation on day t (mm); I [t] is the net irrigation depth on day t that infiltrates the soil (mm).The value of parameter Y was adopted from Siebert and Doll (2010).The irrigation requirement was determined based on the root zone depletion using Eq. 3. The actual irrigation I [t] depended on the extent to which the irrigation requirement was met: where: α was the fraction of the irrigation requirement that was actually met.Following the method as proposed in Hoekstra et al. ( 2009) and also applied by Mekonnen and Hoekstra (2010), Siebert and Doll (2010), 2 scenarios were used, one with α = 0 (no application of irrigation, i.e., rain-fed conditions) and the other with α = 1 (full irrigation).
The measures of green-water evapotranspiration were derived from the minimum values between total crop ET and effective precipitation (Hoekstra et al., 2011).The green-water footprint of vegetable crop production was calculated as a total of the rainwater evaporated from the field during the crop growing period, given as the green component of crop water use (WF green ) divided by the yield t/ha (m 3 /ha).The WF blue of vegetable crop production was calculated as the blue component of crop water use, i.e., water from ground or open body surfaces such as rivers and lakes, following a procedure in the manual by Hoekstra et al. (2011).
The WF grey of vegetable crop production was calculated by quantifying the volume of water needed to assimilate the fertilizers that reach ground or surface water.Nutrients leaching from agricultural fields are the main cause of non-point source pollution of surface and subsurface water bodies.In this study, quantification was done for the grey-water footprint related to nitrogen use only.The WF grey m 3 /t was calculated by multiplying the fraction of nitrogen that leached (θ, %) by the local nitrogen application rate (AR, kg/ha) and dividing this by the difference between the maximum acceptable concentration of nitrogen (C max , kg/m 3 ) and the natural concentration of nitrogen in the receiving water body (C nat , kg/m 3 ) and by the actual individual vegetable crop yields (Y a ) as given in Eq. 4.
The grey-water component was calculated based on the application of nitrogen (N) fertilizer only to the vegetable crop field.The average N fertilizer rates applied to the vegetables were obtained from the agricultural extension officers in the Department of Agriculture.The nitrate leaching fraction was assumed to be 10% of the applied fertilizer rate.The natural concentration of N in the receiving water body was assumed to be zero (Tredoux et al., 2009).Only N fertilizer use was incorporated into the calculations of the WF grey , as N is the most critical pollutant with the greatest application rate (Hoekstra, 2009).Due to unavailability of data, only the recommended fertilizer application rates were used for grey-water footprint estimation at all three SHIS (Table 3).

Economic water productivity (EWP)
Economic water productivity (EWP) (in ZAR/m 3 ) represents the economic value of farm output per unit of water consumed and is calculated as the average producer price for the period 2000-2013 (in ZAR/t).The market price of the vegetable crops was obtained from the abstract of agricultural statistics (DAFF, 2013).The following equation was used to calculate the EWP within the SIS.
where: EWP is the economic water productivity, the summation is done over the years considered from 2000-2013.WF green and blue is the summation of blue-and green-water footprints.The EWP calculations were done following studies done by Schyns and Hoekstra (2014).

RESULTS
It was assumed that farmers in the three SHISs got the same average yield and used the recommended planting dates for each crop (Table 1).This was because year-on-year yield and planting date data were not available for each of the three SHIS for the period 2000-2013.Yield averages and planting dates for the five vegetable crops used in the study were obtained from the extension officers' in the Department of Agriculture.Crop coefficients of different vegetable crops were taken from Allen et al. (1998).

Blue water evapotranspiration
A general increase in ET blue from August to December with a decrease in January was noted.The total average ET blue for ZIS for the entire cabbage growing season was 214.1 mm (Fig. 1).At TFIS the highest average monthly ET blue for cabbage, i.e., 47.2 mm and occurred in December, the lowest was 21.8 mm in August.The total average ET blue for ZFIS in the whole cabbage growing season was 237.5 mm and was the highest among the three schemes.TIS had lowest total average ET blue of 145.3 mm.The highest monthly ET blue (51.2 mm) occurred in October whilst the lowest volume, i.e., 2.5 mm occurred in January.Green beans had the least growing period.Average monthly ET blue generally increased from October, peaked in December then dropped in January (Fig. 2).The average total ET blue for the whole green bean growing season varied amongst the schemes.ZIS had the highest amount of ET blue at 103.3 mm; TFIS had 69.9 mm and the lowest was for TIS with an average of 19.2 mm.

Green water evapotranspiration
Generally, ZIS had the highest total ET green for the full period of the potato growing season (137 mm).TFIS had 74.1 mm while the TIS had the lowest evapotranspiration at 36.1 mm.The highest monthly ET green for both TFIS and TIS was in April (Fig. 3).Of the two schemes, TIS had the highest evapotranspiration of 22.8 mm whilst TFIS had 18.7 mm, which was the lowest.The highest monthly ET green for ZIS was observed in August, i.e., 30.6 mm.ZIS, unlike the other two schemes had its lowest ET green in April (18.5 mm).TIS' lowest ET green was 1.1 mm in August and 12.8 mm for TFIS in June.A rise was noted in ET green from the tomato growing season starting from September to January and a decrease was observed in the month of February (Fig. 4).Early in the production of tomato, TIS had the lowest evapotranspiration of 16 mm but it ended up

Blue and green water footprint
The average water footprint per ton of primary crop differs significantly among crops and across production regions.The WF blue for cabbage production was highest at TFIS, i.e., 79.2 m 3 .At ZIS the WF blue was 71.4 m 3 /t while the lowest amount water was being used at TIS (48.4 m 3 /t).Green beans had the highest WF blue of all the crops in the three SHISs with TIS and ZIS having averages of 1280 m 3 /t and 688.6 m 3 /t respectively, while at TFIS the average is 466 m 3 /t (Table 2).In potato

Figure 4 Green water evapotranspiration for potato production (mm) at Zanyokwe Irrigation Scheme (ZIS), Tugela Ferry Irrigation Scheme (TFIS)
and Thabina Irrigation Scheme (TIS) (period: 2000-2013) production, TFIS used more blue water with an average of 205.2 m 3 /t as compared to 251.7 m 3 /t and 77.7 m 3 /t observed at TIS and ZIS respectively.The lowest amount of blue water calculated at TIS was in the production of tomato with an average of 44.6 m 3 /t.Generally, the average total blue water footprint for production of the five vegetable crops at ZIS is 1 053.7 m 3 /t; at TFIS its 1040.7 m 3 /t while the TIS had the highest total average blue water footprint 1 743.8 m 3 /t.The WF green for green beans at TFIS was 1914 m 3 /t followed by TIS with 1 882.7 m 3 /t and 1 346 m 3 /t at ZIS.Within ZIS, potatoes had the lowest WF green (152.2 m 3 /t) and the highest being green beans.The same scenario was observed at both TFIS and TIS, where potatoes had the lowest WF green and green beans had the highest (Table 2).The study showed that TIS's WF green for cabbage production was 137.7 m 3 /t, while TFIS had 138.3 m 3 /t.The sum of the WF green at the three SHIS revealed that TFIS and TIS used almost the same green water in the production of the five vegetable crops, i.e., 2 561.3 m 3 /t and 2 556.6 m 3 /t respectively, while ZIS had lowest total (2 009 m 3 /t).Amongst the crops green beans had the highest percentage of WF contribution at all the schemes, 62% at ZIS, 68% at TIS and 61% at TFIS (Fig. 5).Cabbage had the lowest percentage contribution in all the three schemes.Green WF had the highest percentage of the water use at ZIS (50.5%), blue water 26.5% while grey water constituted 22.9%.At TFIS blue green and grey water use was 23.1%, 56.7% and 20.2% respectively.Forty-nine (49) of the water footprints at TIS were attributed to green WF, 33.5% to blue and 17.5% to grey WF.

Economic water productivity
Economic water productivity was observed to be lowest in the vegetable crops that had highest blue plus green water in the period 2000-2013.Cabbage had the highest economic water productivity at all the three irrigation schemes, i.e., ZAR 11.02, ZAR 9.16 and ZAR10.70 at ZIS, TFIS and TIS, respectively (Fig. 6).Likewise green beans which had the highest blue plus green water at all three irrigation schemes had the lowest economic water productivity, ZAR 0.98 at ZIS, ZAR 0.84 at TIS. Vegetable production had more value per drop at ZIS then TIS and lastly at TFIS in all the five crops.

Grey water footprint
It was observed that green beans had the highest WF grey , i.e., 373 m 3 /t and the lowest was cabbage with 37 m 3 /t (Table 2).Potato, spinach and tomatoes had 156 m 3 /t, 214 m 3 /t and 132 m 3 /t respectively.

DISCUSSION
Blue and green water evapotranspiration ET demand of a crop is a measure of how much water can be consumed via soil evaporation and plant transpiration assuming that plant-available water is adequate (Andales et al., 2011).ET blue is mostly irrigation water, which is often used to offset the impact of rainfall variability on crop yield and to reduce the risk associated with weather variability (Guerra et al., 2005).Generally, ET blue was lower than ET green at all the three sites (Figs 1-4), possibly because the vegetable crops were produced under rain-fed conditions.Irrigation is usually used as a supplement to shortages during the growing season.
Different vegetable crops have different blue and green ET at the three sites; this may be attributed to the differences in weather and environmental characteristics in the three locations.According to Dukes et al. (2012), the needs for ET are primarily dependent on soil type, crop growth and development which are closely related to climatic demands.The ET demand varies from day to day depending on crop growth stage and weather variables (Andales et al., 2011).Vegetable crops have a high requirement for water during the growth season, which could be the reason for the high ET in the three SHISs.The differences in ET of crops at the same SHIS could be due to the physiology and biology of the crops; for example, some crops have inherently high transpiration than others.For example at ZIS the ET blue for cabbage was the highest at 328.1 mm, tomatoes had 311.1 mm and the lowest ET blue at the same irrigation scheme was observed in green beans, i.e. 201.9 mm.The differences can also be seen in among the SIS were for spinach ET blue values of 166.4 mm, 144.5mm and 183.5 mm were observed at ZIS, TFIS and TIS respectively.Krug et al. (2002), noted that evapotranspiration and crop coefficient varies in the course of the season because morphological and eco-physiological characteristics of the crop do change over time.

Water footprint of growing vegetable crops
The study finds that WF green is dominant, making up 50.5% at ZIS, 56.7% at TFIS and 49% at TIS.This highlights the fact that green water is a very important resource for food production at the SIS and globally.This was the case in studies done by Schyns and Hoekstra, ( 2014) and Mekonnen and Hoekstra (2011) where WF green was 77% in Morocco and a global WF green of 78% respectively.Results from the current study are relatively higher than those of similar studies done in other countries.Chapagain and Orr (2009) found the national virtual water contents of Spanish tomatoes were 60.5 m 3 /t for WFblue and 13.6 m 3 /t for WF green.Mekonnen and Hoekstra (2011) calculated the global average blue and green WF of tomatoes to be 63 m 3 /t and 108 m 3 /t, respectively.Aldaya et al. ( 2010), calculated Italian tomato production to have 35 m 3 /t WF green and 60 m 3 /t WF blue .Mekonnen and Hoekstra (2011), the global WF blue and WF green of cabbage production was 181 m 3 /t and 26 m 3 /t respectively, these are considerably lower than the results observed in this study (Table 2).One possible reason could be the low yields that smallholder farmers get and the difference in planting dates (Table 1), the lower the yields the higher the water footprint everything else remaining constant.This was also suggested by Mekonnen and Hoekstra (2011), were they observed that crops with a higher yield or a large fraction of crop biomass that is harvested generally have a smaller water footprint per ton than crops with a low yield or a small fraction of crop biomass harvested.Contrary to the above findings, the results of this study can possibly be attributed to the differences in production and weather characteristics as alluded to by Gerbens- Leenes et al. (2008).Ridoutt and Pfister (2013) and Mekonnen and Hoekstra (2011) also highlighted that the potential environmental impacts related to water use are different from one location to another.Jordaan and Grove (2012) observed that blue water required to produce cabbages and carrots at ZIS was 763 m 3 /t and 273 m 3 /t respectively.These results are significantly different from the WF blue observed in this study.The difference can possibly be attributed to a variety of causes, including: type of model, spatial resolution, the period considered and data regarding cultivated and irrigated areas, growing periods, crop parameters.The SAPWAT programme used in the above study uses climatic data from 1957 to 1999 (Van Heerden and Crosby, 2008).Another important factor is the change in the length of the growing period may notably vary the crop water use and thereafter the green and blue water footprint obtained (Chico et al., 2010).The differences indicate that the methodology applied is sensitive to input of climatic data and assumptions concerning the start of the growing season (Gerbens-Leenes et al., 2008).
To help reduce the water footprint within the smallholder irrigation schemes, it is important to take the timing of crop water demands with respect to natural water availability into account in deciding which crops or crop varieties to grow (Schyns and Hoekstra, 2014).Other ways of helping reduce the WF within the SIS would be to introduce measures like mulching, conservation tillage, and small-scale water harvesting techniques which increase infiltration and reduce runoff and evaporation.Although it is mostly the farmers choice on which vegetable crops to grow, they can however be advised especially through extension agents to focus more on crops that give a high income per drop of water.The choice of a vegetable, variety length of growing season largely determines the final WF.For example at ZIS rainfall is erratic and low throughout the year, short season varieties can be grown especially in the summer season where the amounts rainfall is mostly concentrated (Fig. 1).At TIS they receive high rainfall during the summer hence it is advisable to produce vegetable during this period (Nov-March).According to the results from the current study it is advisable to shift green beans production from TIS to both ZIS and TFIS while tomato production can be shifted to ZIS and TIS were it has a higher EWP value (Fig. 6).
There are a number of uncertainties and limitations to the study of WF.According to Chenoweth et al. (2014), a variety of methodologies have been developed for water foot printing which differ with respect to how they deal with different forms of water use, this creates some understandable scepticism and hesitance when it comes to interpreting the meaning and relevance of different water footprint estimates.Estimates of WF in this study were largely affected by the unavailability of data as most smallholder farmers do not keep records of the actual yields and chemicals used during a particular season.Different farmers manage their crops differently hence there is need to do a hands on study where the farmers are monitored and real time data collected for purposes of accurately predicting the WF of each crop.Usually smallholder farmers do not sell all their crops to the market, some they use for home consumption while some is sold at the farm at relatively lower prices than the market.This presents a challenge because no actual yield data and price is available for a proper economic water productivity analysis.

SUMMMARY AND CONCLUSIONS
The study shows that the water footprint of vegetable production varies within different irrigation schemes in different regions for the period 2000-2013.The grey water footprint is relatively small and was assumed to be the same at all the irrigation schemes, this is because only nitrogen fertilizer leaching from the SHIS was considered, leaving out other pollutants such as pesticides and herbicides.This was done due to the unavailability of the actual application rates of the other chemicals.Planting period used in the study were obtained from the Department of Agriculture and literature reviews.It was, however, noted that there is variation of the same crop in the planting dates even within the same schemes and across the country.The choice of the planting and harvesting dates out of these ranges obviously influences the final crop water footprint estimate.Blue water is only required when the green water store in the soil runs short, so it is important that farmers synchronise their planting dates with the rainy seasons so that they can reduce on the use of blue water and make use of the green water in vegetable production.
TIS had the highest WF followed by TFIS whilst ZIS had the lowest.There are large differences in WF in SHIS that are caused by a number of factors, chief among them are weather and production practices.More water is used in crop production within SHISs in SA compared to other countries as evidenced by the high WF observed in this study.This can be an indicator that the yield is not proportional to the drop of water, hence a need to increase crop output.Thabina Irrigation Scheme had the highest WF blue in green beans and potato and the least in tomatoes.Green beans WF blue at observed TIS was 1280 m 3 /t, and ZIS had 688.6 m 3 /t while the lowest was observed at TFIS (466 m 3 /t).TFIS had the highest WF blue in cabbage, spinach and tomatoes.However, the results show that there is more WF green in vegetable production at all the three sites except for potato at both TFIS and TIS.WF green is highest in green beans compared to the other crops in all the SHISs.The results also show that in potato production, more blue water was used with the highest WF green recorded at ZIS. Spinach and tomatoes had the highest WF green at TIS.An improvement in the production practices and ultimate yield is necessary to reduce the WF for improving water use efficiency.

Figure 7
Figure 7Contribution of different crops to the total water footprint of crop production (mm) at three irrigation schemes in South Africa(period: 2000-2013)

TABLE 1 Average yields of vegetable crops grown in smallholder irrigation schemes (SHIS) in South Africa Crop Yield (t/ha) Month planted
The average of total ET green of the whole season growing tomatoes TIS had the highest average of 443.2 mm followed by TFIS having 418.5 mm and ZIS had the lowest average of 311.1 mm.The highest monthly ET green was in November for TIS it had 118 mm, TFIS had 111.7 mm in December.ZIS had the highest monthly ET green in January having 78.6 mm.The lowest monthly ET green for TIS was 16 mm, TFIS had 19 mm and ZIS had 12.6 mm.