Data collection and treatments

The research was carried out at 2,660 m a.s.l., temperature 13.9ºC and 5°31'53.98'' N and 73°29'15.7'' (Samaca, Boyaca, Colombia).The hybrid Osaka F1 yellow onion with a short and early day was used. The evaluations were carried out from planting to harvest (first half 2021) in Fluventic Haplustepts soil (^{USDA, 2014}). The treatments consisted of two irrigation systems, sprinkler and drip, with distances between sprinklers of 12 m between lines and 9 m between emitters, and drip, five tapes per bed with a distance between lines of 0.4 m and between drippers 0.1 m.

To determine the water footprint, 1×1×0.35 m lysimeters were used, with a storage cabin and water extraction outlet. The extraction was done weekly using suction pumps, together with 50 g of soil to determine gravimetric humidity. Chemical analysis (N, P, K, Ca, Mg and K) and physical analysis of soils were carried out: apparent density (g cc^-1) (known volume ring) and real (pycnometer), moisture retention curves (pressure cookers, 2.5 cm rings) at 0.33 and 15 bars (field capacity, permanent wilting point), hydraulic conductivity (cm h^-1) (constant head hydraulic conductivity equipment, 5 cm rings), and physicochemical analysis of irrigation water.

A portable weather station was also used for the collection of precipitation, evapotranspiration and temperature, and the Cropwat application for the determination of evapotranspiration (EV) and daily water balances, every 10 d (1 dec) and monthly (^{Allen et al., 2006}; ^{Romero et al., 2016}). Precipitation information over a 30-year period from the IDEAM Villa Carmen station (Samaca) was used.

Weekly data were collected on lysimeter leachate (mm), applied irrigation sheets (mm h^-1), gravimetric moisture content (%), precipitation (mm) and reference evapotranspiration from the information generated by the portable weather station. Cropwat was used to determine crop evapotranspiration and effective precipitation (ETc [mm dec^-1]).

During the crop cycle, pest and disease control was ensured. Edaphic fertilization was applied in each treatment based on the chemical analysis of soils and fertigation formula (15 d after emergence until day 60) and maturation (begins day 61 up to 30 d before harvest) (Tab. 1).

Components of the water footprint (WF)

In line with ^{Hoekstra (2017)}, WH_blue is the water applied by artificial means for crop production; it was calculated from crop water use (CWU [m³ ha^-1]) divided by crop yield (Y [t ha^-1 ], Eq. 1). The CWU was determined from the daily evapotranspiration (mm d^-1) during the evaluation period (Eq. 2).

where the EVT_blue is the total evapotranspiration of blue water, 10 the conversion factor to mm, LPC is the cultivation period from the day of transplantation to harvest.

Since the WF_green corresponds to the contributions of water from rainfall, it is estimated by taking the results of effective precipitation (PRECe) using Cropwat as a reference. The effective precipitation is that part of the precipitation stored in the soil and potentially available to satisfy the needs of the crop. Most of the time the effective precipitation is less than the precipitation because plants cannot appropriate all the water, due to surface runoff and percolation (Eq. 3).

where PRECe was the effective precipitation and y the yield.

Finally, the WF_gray is the volume of water necessary to assimilate the pollutants (the pollutants generally referenced in agricultural systems are fertilization and pesticide residues) generated up to the limits allowed by current legislation (^{IDEAM, 2019}) and is obtained by dividing the load polluting by difference between the maximum permissible level of a pollutant and the natural concentration of it, divided by yield (^{Franke et al.,2013}) (Eq. 4).

where α was the fraction of leaching that reaches fresh water, with 10% assumed for all fertilizers (^{Chapagain and Hoekstra, 2011}), AR the amount of fertilizers applied (Tab. 1), Cmax the maximum acceptable concentration of leachate from fertilizers and reference values under Resolution 1575 (^{MAVDT, 2007}) for chemical substances related to pesticides and fertilizers (for our case, a maximum value for nitrates (NO₃) of 10 mg L^-1 was assumed), and Cnat, the natural concentration of pollutants (value of zero (0) was assumed).

Edaphoclimatic characteristics

Climate monitoring allows adequate management of different agronomic activities to achieve a productive and sustainable model. The average temperature fluctuated between 13.5 and 14.4°C, with an average of 13.9°C, for February, 14.3°C for April and May, and between 13.5 and 13.8°C for the remaining months, typical of the cold climate of the area. The accumulated precipitation (565.4 mm) was lower than the accumulated evapotranspiration (919.4 mm), and the months with the highest precipitation (>100 mm), and the lowest evapotranspiration (89.5 and 97.4 mm) were March and May, although April registered 96 mm of precipitation. The monthly evapotranspiration was greater than the precipitation in the months of January, February, April, June, July, August and September, resulting in a water deficit (Fig. 1).

Figure 1. Precipitation, evapotranspiration and average temperature of the study area. 

Water management must take full advantage of the weather conditions according to the results and the actual situation (^{Gao et al., 2021}). Some studies have shown that the water requirement of the crop not only depends on its particular characteristics but also on the moment of the crop cycle, the environmental conditions and the planting location (^{Zhang et al., 2021}).

The WF is mainly determined by temperature and precipitation (^{Zhang et al., 2015}). To evaluate the efficiency and productivity of water, it is necessary to understand that its use is directly related to the water requirement of the crop, which can be supplied by water stored from precipitation in the soil as moisture (green water) or by irrigation (blue water) (^{Adeoti et al., 2021}).

The main properties of the soil (Tab. 2) must be taken into account for the dynamic management of water through the efficient management of irrigation and the use of rainwater.

The predominant texture was clayey (>65% content), which is reflected in a low hydraulic conductivity (0.09 cm h^-1). The percentage of pores is adequate (49.1%); nevertheless, the distribution between macro (1.5%), meso (14.9%) and micropores (32.7%) is not favorable for drainage, due to the high clay content, since the movement, content and availability of water are subject to texture, structure, porosity, depth and organic matter content (^{Valenzuela and Torrente, 2010}). Onion cultivation does not require deep soil, with 40-60 cm being sufficient, as long as there is good drainage, and develops in a pH range of 6.0 to 7.5 (does not tolerate acidity) and medium textures (^{Ruiz et al., 2013}; ^{Adeoti et al., 2021}).

The pore space of the soil is sensitive to natural and anthropogenic impacts; any disturbance of the superficial soil layer (0-30 cm) by rain or tillage causes changes in its structure, affecting some soil hydraulic properties. Excessive tillage and loss of soil organic matter often lead to a reduction in the infiltration rate due to loss of surface porosity. When the intensity of the rain is greater than the infiltration rate, runoff ensues (^{Shaxson and Barber, 2005}), hence the importance of the hydrodynamic parameters of the soil (^{Castro et al., 2009}).

To calculate the water footprint, a summary of the data estimated during the onion bulb crop cycle is presented in Tab. 3.

Blue water footprint

The WF_blue not only includes the water evapotranspired and incorporated into the product, but also the water that evaporates during its storage (artificial water tanks), transport (open channels), treatment (in case it is not suitable for irrigation) and application efficiency, being a completely anthropic decision (^{Allen et al., 2006}; ^{Romero et al., 2016}). A greater blue water footprint was found in the drip irrigation system (78.7 m³ t^-1) compared to sprinkler (75.7 m³ t^-1) (Fig. 2).

Figure 2. WF_blue, WF_green and WF_grayin the cultivation of Allium cepa for sprinkler and drip irrigation systems 

Green water footprint

The use of water from precipitation together with the properties of the soil will determine the green water footprint. Higher consumption was found in the sprinkler irrigation system (67.53 m³ t^-1) versus drip irrigation (65.3 m³ t^-1) (Fig. 2).

Gray water footprint

According to ^{IDEAM (2019)}, leached water from agricultural activity contains intermediate compounds of the original molecules, including toxic cations, mineral nutrients, herbicides, and pesticides that would not have been mobilized without irrigation. Assuming that 20% of the amount of nitrogen applied during fertilization would contaminate the leached waters, mainly with nitrate, the application of nitrogen fertilization in sprinkler irrigation (147.8 kg ha^-1) was 63.2% of that applied under drip irrigation of (233.7 kg ha^-1)(Tab. 1), which created a larger gray footprint (52.4 m³ t^-1) with respect to sprinkler irrigation (31.3 m³ t^-1) (Fig. 2).

For onion cultivation under the conditions of Samaca (Colombia) and using the most common irrigation systems (Fig. 2), WF for sprinkler irrigation was 174.5 m³ t^-1 with a yield of 56.02 t ha^-1 compared to drip irrigation of 196.4 m³ t^-1 with a yield of 57.9 t ha^-1.

The WF_green and WF_blue were similar in the two irrigation systems (sprinkler 143.1 m³ t^-1 and drop 144 m³ t^-1). ^{IDEAM (2019)} mentions that transitory crops represent 18% of the total agricultural water demand, 13% of the total WF_blue and 3% of the total WF_green. In addition, ^{Léllis et al. (2022)}, mention that in the case of WF_green, the amount and distribution of precipitation condition WF_blue, because in areas where rainfall supplies all or most of the crop's water requirements, this indicator is low.

^{Zhang et al. (2021)} propose efficient water management agricultural practices to reduce the consumption of irrigation water, increase crop yields and develop sustainable agriculture to guarantee food security. Improving the rate of use of green water or reducing the use of blue water in agricultural planning can save irrigation water without reducing yield (^{Su et al., 2014}).

Thus, for example, a greater WF_graywas found in drip irrigation (52.41 m³ t^-1) compared to that of sprinkler (31.29 m³ t^-1) due to fertigation, which applies a greater amount of nitrogen.

According to ^{Mekonnen and Hoekstra (2011)}, the world average of the total water footprint for vegetables is 322 m³ t^-1, (WF_blue43, WF_green 194 and WF_gray 85 m³ t^-1) and for onion specifically 271 m³ t^-1 (WF_blue 44, WF_green 176 and WF_gray 51 m³ t^-1). ^{Peñaloza-Sánchezet al. (2020)} report, for Mexican conditions, a WF of the onion crop for the spring/summer cycle of 578.65 m³ t^-1 (WF_blue263, WF_green293.3, and WF_gray 22.9 m³ t^-1) with a low yield of 7.69 t ha^-1, contrary to the autumn/winter cycle withWF of 249.7 (WF_blue 53,WF_green 176.4, and WF_gray 20.1 m³ t ^-1), with a yield of 20.4 t ha^-1.The authors mention that the gray water footprint was underestimated because only the use of fertilizers and not pesticides was considered. Our results for sprinkler and drip irrigation systems (174.47 and 196.41 m³ t^-1) were lower, while our results for WF_gray in sprinkler (31.3 m³ t^-1) and drip (52.4 m³ t^-1) were higher.

In Colombia, the WF of agricultural indicator with the highest participation is WF_green, due to the high rainfall that implies a lower requirement for irrigation water. This means that the cost of using water for agriculture is lower than in other countries, where agriculture depends on infrastructure and technologies for crop irrigation (^{IDEAM, 2023}).

^{Castañeda and Ramírez (2016)}, for Colombia, report 164.59 m³ t^-1 for WF_blue and 324.07 m³ t^-1 for WF_green, higher than our results where the WF for the sprinkler irrigation system was 174.47 m³ t^-1 and for drip 196.41 m³ t^-1. The Instituto de Hidrología, Meteorología y Estudios Ambientales (^{IDEAM, 2019}) estimated a national WF_blueof 8.5 Mm³ year^-1 and WF_green of 34.6 Mm³ year^-1 for the bulb onion crop for the year 2018, using the statistics on harvested area (12,695 ha), production volume (287,959 t) and average yield (22.68 tha^-1) from the ^{Ministry of Agriculture and Rural Development (2019)}. According to these estimates, WF_blue was 29.51 m³ t^-1 and WF_green 120.15 m³ t^-1, which are lower than our results for WF_blue (sprinkler 75.7 m³ t^-1and drip 78.7 m³ t^-1), higher for WF_green (sprinkler 67.5 m³ t^-1 and drip 65.3 m³ t^-1) and lower for yield (sprinkler 56.0 t ha^-1 and drip 57.9 t ha^-1). Other results on onion water footprint are listed in table 4.

The information on the use of water by agricultural activity is limited for the Colombian context. The water footprint is a powerful instrument that not only quantifies the volume of water used but also allows for better multi-objective planning for saving water, evaluating new technologies, comparing production systems, evaluating different edaphoclimatic moments, reducing wastewater, finding the best economic gain, and seeking the greatest social impact (generated jobs), among others. The pressure on water will increase due to changes in ecosystems, climate change and the constant increase in the demand for food due to the increase in the population (^{Arévalo et al., 2011}; ^{Ríos et al., 2018}).