**1. Introduction**

Climate change is becoming more and more relevant for the agricultural sector, and its most catastrophic events are causing huge losses in crop productivity. A chance to effectively react to such extreme events and adapt to a constantly changing scenario is offered by innovative Agri-Tech technologies. Unfortunately, such solutions are mostly adopted only in large-scale production industries, whereas smaller companies often lack technological expertise, aptitude for innovation, or enough financial resources for the initial investments. The goal of this project is to study and develop a specific solution capable of monitoring the health of a crop field in an automatic and non-invasive way and that does not need a high economic effort nor advanced technical skills to be applied. This study focuses on Italian vineyards since they are usually part of small–medium-sized companies, forming a fragmented economic context based on traditional techniques and with a generally low diffusion of technological innovation. Vineyards are now facing the impact of climate change and, due to their peculiar characteristics, are more prone to its consequences. A further objective is to foster the integration of the traditional know-how

**Citation:** Brach del Prever, P.; Balducci, G.; Ballestra, A.; Ghiglione, C.; Mascheretti, L.; Molinari, M.; Nicoletti, G.; Carvelli, V.; Corbari, C.; Invernizzi, S.; et al. Automatic and Non-Invasive Monitoring of Water Stress in Vineyards. *Environ. Sci. Proc.* **2023**, *25*, 79. https://doi.org/ 10.3390/ECWS-7-14164

Academic Editor: Lampros Vasiliades

Published: 14 March 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of older farmers with the technological expertise of younger ones, thus increasing the resilience towards future environmental scenarios.

One effective parameter to monitor the health status of vineyards is the Crop Water Stress Index (CWSI), a quantitative estimation of a plant's need for water. In detail, water availability is a key parameter influencing the quantity and quality of vineyard production. The knowledge of water stress is an important feature for controlling the optimal water conditions needed to achieve the desired characteristics of production. Hence, the aim of this study is to design and implement an intelligent autonomous system to measure the spatial and temporal distribution of CWSI on a crop field, offering a valuable instrument to have precise insight into the health of vineyards and to schedule targeted interventions to recover optimal values. Furthermore, geospatial and historical analysis of CWSI can be performed to gain insight into the characteristics of the crop field and infer connections with the productivity of the harvest.

Some solutions for the monitoring of vine water stress are available, but their application is still at an early stage and the technology is not ye<sup>t</sup> refined [1]. To the best of our knowledge, no water resource optimization system, via active and adaptive crop monitoring, is available on the market at affordable prices.

The potentialities of the proposed solution have been tested in a data collection session in a vineyard owned by "Azienda Agricola Balladore Pallieri" (Asti, Italy) through a low-cost prototype. For this purpose, miniaturized sensors have been assembled on a Raspberry Pi. This on-field campaign helped us to verify the validity of the concept and its implementation to collect climatic parameters and elaborate the data into real CWSI heatmaps. Moreover, it contributed to highlighting its advantages and the further improvements of the design.

The system provides almost real-time information on the hydration condition of the field at near-real-time with a grea<sup>t</sup> level of geographic detail, enabling wise water managemen<sup>t</sup> and boosting plant care. Thanks to the CWSI maps of the field, it is possible to focus watering only on the areas of the vineyard needing water, thus reducing water use.

The system fits well with the current state of society, responding to one of the most urgen<sup>t</sup> challenges: climate change. The proposed solution increases the resilience of the economic system to climate change as it improves the farmer's ability to control and optimize the water usage in their fields. Moreover, the flexibility of the proposed sensing system and its software suite allows for adaptive and evolving technology. Indeed, the project could successively be adapted to fit the needs of other kinds of crops, thus expanding its market and impact on the agriculture industry.

### **2. Background, Parameters, and Methods**

The Crop Water Stress Index [2,3], or CWSI, is chosen as the key vegetation index in this work as it is strongly related to the water status of the crop. It measures the ability of the plant to exploit the available water in the soil. It is related to the health of the plant because it describes the ability to successfully bring water from the roots to the leaves and fruit. Instead of measuring it directly through the evaluation of stomatal pressure, which would require complex instrumentation, it can be evaluated through climatic and atmospheric data in the area surrounding the plant as it is defined as:

$$\text{CWSI} = 1 - \frac{\text{ET}}{\text{PET}} \tag{1}$$

where ET and PET are measures of evapotranspiration potential, closely linked to the water potential of the plant. A way to calculate this index based on meteorological data was proposed by [3]:

$$\text{CWSI} = \frac{\gamma \left(1 + \frac{r\_a}{r\_a}\right) - \gamma}{\Delta + \gamma \left(1 + \frac{r\_a}{r\_a}\right)} \tag{2}$$

$$\frac{r\_c}{r\_a} = \frac{\frac{\gamma r\_a R\_u}{\rho c\_p} - \frac{T\_C - T\_A}{\Delta + \gamma} - (p\_{v,sat}(T\_C) - p\_v)}{\gamma \left[ (T\_C - T\_A) - \frac{r\_a R\_u}{\rho c\_p} \right]} \tag{3}$$

where:


The value of CWSI ranges from 0 to 1, where 0 stands for a healthy plant with no water stress, while 1 represents a critical situation of water deprivation. CWSI values can be subdivided into four ranges of plant conditions (Table 1), exploiting the correlation between CWSI and Leaf Water Potential available in the literature [4].

**Table 1.** CWSI values and corresponding water stress of the plant.


### *2.1. Satellite Data*

The input meteorological data were all taken from the European Database Copernicus, which provides time series with hourly precision over specific regions of the earth, apart from the canopy temperature, which was downloaded directly from Landsat Satellites through Google Earth Engine as the Land Surface Temperature (LST) parameter. Whereas Copernicus offers hourly data every day, Landsat Satellite only detects the data over a specific route once every 15 days, so we filtered the Copernicus data only for the days where we could also have the LST to ensure the best compatibility of the data coming from two different sources.

### *2.2. Design of the System*

The system has two main parts: the sensing part collects information on the vines and the vineyard status, while the analysis part computes the Crop Water Stress Index (CWSI) and generates maps that can be used to monitor the current status of the vineyard in detail, as well as to make predictions about the future development of the grapes and the plants.

The sensing subsystem is a battery-powered module composed of specific sensors interfaced with a Raspberry Pi 4 Model B. To collect climatic parameters and elaborate the data into real CWSI heatmaps, the following sensors were adopted (Figure 1a):


**Figure 1.** (**a**) Simplified scheme of the raspberry box; (**b**) Bird view of the selected field of "Azienda Agricola Balladore Pallieri", A, B, C and D (small area south of Sector B) show waypoints.

### *2.3. Data Collection*

In order to verify the validity of the proposed concept, an on-field data collection session in a vineyard owned by "Azienda Agricola Balladore Pallieri" in Calosso (Asti, Italy, see Figure 1b) was carried out on 4 June 2022. The location is characterized by a mild climate, scarcity of droughts, and strong hailstorms. The test was carried out in late spring to investigate the status of mature vines, and the good weather conditions ensured a more reliable measurement of irradiation.

Precise waypoints were chosen in the crop to create a network of measurements roughly every 10 m, to increase the resolution of the CWSI map with respect to the satellite measurements (Figure 1b). The waypoints were divided into three sections in different zones of the crop (A, B, C, D) to investigate the side-by-side areas variation of CWSI linked to changes in slope, which lead to different water drainage.
