A Practical Look at the First Week

When we started working with a grower in the eastern zone of Mendoza, the goal seemed simple: measure soil moisture for seven days and adjust irrigation accordingly. Reality, as usual, was less straightforward. This is an account of that first week, with the numbers, the doubts, and the adjustments that came up along the way.

The starting point was choosing three plots with different histories: one with newly established alfalfa, another with grapevines in production, and a third that had been left fallow. The idea was not to compare them against each other, but to understand how each responded to the same irrigation depth. To do this, we installed moisture sensors at 20, 40, and 60 centimeters deep, and recorded readings every four hours.

The first surprise came on the second day. The alfalfa plot showed a higher-than-expected surface moisture reading, but at 40 centimeters the value dropped sharply. The initial diagnosis pointed to a compaction problem, but when we reviewed the tillage history, we found that the seedbed had been prepared with too many passes of the disc harrow. The solution was not to change the irrigation, but to reschedule the next ripping operation.

In the vineyard plot, the behavior was different. Moisture remained stable during the day and only dropped after six in the evening, when the plant resumed its activity. That seemingly minor detail led us to shift the irrigation shift two hours later to take advantage of the window of greatest demand. The change was subtle, but by the end of the week, water use per vine had dropped by 12% without affecting crop vigor.

The fallow plot, for its part, served as a control. With no active vegetation, moisture remained almost constant over the seven days, except for a slight loss at the surface due to direct evaporation. That data helped us calibrate the sensors and rule out that the variations observed in the other plots were due to a measurement error.

Not everything was easy. On the third day, a failure in the data logger in the alfalfa plot left us without readings for nearly ten hours. Instead of repeating the measurement, we decided to take manual gravimetric samples at the same points and compare them with the values from the previous day. The difference was minimal, confirming that the overall trend held.

The week ended with three practical conclusions. First, measurement depth matters as much as frequency: the data at 20 centimeters told one story, but the data at 60 centimeters told another. Second, the management context—tillage, crop history, time of day—weighs more than any isolated reading. And third, a monitoring plan is useless if it is not accompanied by a clear record of decisions, even if those decisions are as simple as moving an irrigation shift by two hours.

For those thinking about doing something similar, the advice is to start small. A single plot, two or three sensors, and one week of data are enough to understand the general behavior. After that, you can scale up. What you should not do is jump straight into a complex system without first validating that the readings reflect what is actually happening in the field.

The next stage will be to repeat the monitoring after a rainfall event, to see how infiltration behaves in each plot. We also want to add a fourth measurement point in an area with a history of waterlogging, which we left pending due to access issues. If this first week taught us anything, it is that useful data appears when you take the time to look at it with context.