Synthetic Aperture Radar in Vineyard Monitoring: Examples, Demonstrations, and Future Perspectives

Abstract

Synthetic aperture radar (SAR) technology has emerged as a pivotal tool in viticulture, offering unique capabilities for various applications. This study provides a comprehensive overview of the current state-of-the-art applications of SAR in viticulture, highlighting its significance in addressing key challenges and enhancing viticultural practices. The historical evolution and motivations behind SAR technology are also provided, along with a demonstration of its applications within viticulture, showcasing its effectiveness in various aspects of vineyard management, including delineating vineyard boundaries, assessing grapevine health, and optimizing irrigation strategies. Furthermore, future perspectives and trends in SAR applications in viticulture are discussed, including advancements in SAR technology, integration with other remote sensing techniques, and the potential for enhanced data analytics and decision support systems. Through this article, a comprehensive understanding of the role of SAR in viticulture is provided, along with inspiration for future research endeavors in this rapidly evolving field, contributing to the sustainable development and optimization of vineyard management practices.

1. Introduction

Remote observation of the Earth from its orbits has a rich history dating back to the mid-20th century. Demonstrating numerous achievements that highlight unparalleled capabilities, remote sensing is now shifting its focus from purely technological advancements to practical applications and the development of more robust processing methods. However, with continuous advancements in remote sensing systems, there is a growing challenge of addressing various unwanted factors, such as atmospheric interference, weather conditions, day/night limitations, topographic effects, and limited sensitivity to certain environmental properties and structures. The technology of synthetic aperture radar (SAR) offers a notable advantage. By utilizing active remote sensing systems, SAR can generate high-resolution imagery regardless of the time of day and weather conditions.

SAR instruments have several shared features despite their diverse designs. A critical advantage of these instruments is their use of active radar systems. Unlike passive instruments, active radars illuminate the Earth’s surface using an artificial source of microwaves emitted by antennas onboard the satellite. This capability ensures operational independence from daylight and atmospheric conditions.

SAR instruments typically operate in the microwave spectrum, particularly in the L-, S-, C-, or X-bands, as detailed in

Table 1. Synthetic aperture radar satellite missions and details on their operational characteristics.

Satellite	Country	Agency	Years	f (GHz)	$\mathit{\lambda }$ (cm)	Band	Pol.	ΔT (Day)
Seasat	USA	JPL	1978–1978	1.27	23.5	L	HH	3
SIR-B	USA	JPL	1984–1984	1.28	23.4	L	HH	-
Magellan 1	USA	JPL	1989–1992	2.39	12.6	S	linear	var
ERS-1	EU	ESA	1991–2000	5.30	5.7	C	VV	35 2
JERS-1	Japan	JAXA	1992–1998	1.27	23.5	L	HH	44
SIR-C/X-SAR	USA	JPL	1994–1994	1.24	24.2	L	SP	1
				5.29	5.7	C	SP	1
				9.60	3.1	X	VV	1
ERS-2	EU	ESA	1995–2011	5.30	5.7	C	VV	35 3
RADARSAT-1	Canada	CSA	1995–2013	5.30	5.7	C	HH	24
SRTM	USA	JPL	2000–2000	5.30	5.7	C	DP	0
				9.60	3.1	X	VV	0
Envisat	EU	ESA	2002–2012	5.33	5.6	C	SP, DP	35
ALOS	Japan	JAXA	2006–2011	1.27	23.6	L	SP, DP, QP	46
RADARSAT-2	Canada	CSA	2007–present	5.30	5.7	C	SP, DP	24
TerraSAR-X	Germany	DLR	2007–present	9.65	3.1	X	SP, DP	11 4
COSMO-SkyMed 5	Italy	CSI	2007–present	9.65	3.1	X	SP, DP, QP	16 6
TanDEM-X	Germany	DLR	2010–present	9.65	3.1	X	SP, DP, QP	11 7
RISAT-1	India	ISRO	2012–2016	5.35	5.6	C	SP, DP, QP	25
HJ-1C	China	CRESDA	2012–present	3.13	9.6	S	SP	31
Kompsat-5	Korea	KARI	2013–present	9.66	3.2	X	SP	28
Sentinel-1A	EU	ESA	2014–present	5.41	5.5	C	SP, DP	12 8
ALOS 2	Japan	JAXA	2014–present	1.26	23.8	L	SP, DP, QP	14
Sentinel-1B	EU	ESA	2016–2021	5.41	5.5	C	HH, VV, DP	12 9
Gaofen-3	China	CNSA	2016–present	5.00	5.5	C	SP, DP, QP	29
TerraSAR-X NG	Germany	DLR	2018–canceled	9.65	3.1	X	SP, DP, QP	2
PAZ	Spain	INTA	2018–present	9.65	3.1	X	SP, DP	11
NOVASAR-1	UK	SSTL	2018–present	3.30	10.0	S	SP, DP	14
SAOCOM 1A	Argentina	CONAE	2018–present	1.27	23.5	L	SP, DP, QD	8 10
RCM 11	Canada	CSA	2019–present	5.41	5.5	C	SP, CP	12 12
SAOCOM 1B	Argentina	CONAE	2020–present	1.27	23.5	L	SP, DP, QD	8 13
CSK 2nd Gen.	Italy	CSI	2022–present	9.65	3.1	X	SP, DP, QD	16 14
NISAR	India, USA	NASA, ISRO	2024–planned	1.20	24.0	L	SP, DP, QD	12
				1.50	12.0	S	SP, DP, QD	12
Biomass	EU	ESA	2024–planned	0.44	70.0	P	SP, CP	25
Sentinel-1C	EU	ESA	2024–planned	5.41	5.5	C	SP, DP	12 *
ALOS-4	Japan	JAXA	2024–planned	1.26	23.8	L	SP, DP, QP	14
Sentinel-1D	EU	ESA	2025–planned	5.41	5.5	C	SP, DP	12 *
TanDEM-L	Germany	DLR	2028–planned	1.27	23.6	L	SP, QP	16
ROSE-L	EU	ESA	2028–planned	1.26	23.8	L	DP, QP	6 *
Harmony	EU	ESA	2029–planned	passive	passive	C		12 *
Sentinel-1 NG	EU	ESA	2032–planned	5.41	5.5	C	SP, DP, QP	<6 *

f, carrier frequency; $\lambda$, wavelength; $\Delta T$, nominal revisit time; Pol., polarization; SP, single polarization (HH/HV/VV/VH); DP, dual polarization (HH + HV/VV + VH); QP, quad polarization (HH + HV + VV + VH); CP, compact polarimetry; ¹ mission to map the surface of Venus; ² valid during period 1992–1993/1995–2000, 3 days in 1991–1992/1993–1994, 168 days in 1994–1995; ³ 24 h tandem with ERS-1; ⁴ minimum 2.5 days with 95% probability; ⁵ constallation of 4 satellites; ⁶ minimum 12 h; ⁷ close formation (250–500 m distance) with TerraSAR-X; ⁸ 12 days in Europe; ⁹ 12 days in Europe, 6 days in constellation with Sentinel 1-A; ¹⁰ minimum 12 h in constellation with COSMO-SkyMed; ¹¹ 3 satellites constellation; ¹² minimum 4 days; ¹³ minimum 12 h in constellation with SAOCOM-1A and COSMO-SkyMed; ¹⁴ minimum 12 h; * constellation missions.

. To accurately capture surface details, radars must adopt a side-looking orientation, which involves angling the radar to obtain a side view of the terrain (refer to Figure 1). The radar emits pulses sideways, perpendicular to its flight direction, illuminating a swath of the ground. As the satellite moves, the radar antenna transmits pulses and simultaneously collects the reflected signals, allowing it to map a strip of land parallel to its flight path. This method, known as the strip-map mode, is the most widely used sensing mode in satellite SAR techniques [1].

There are several other imaging modes available, depending on the satellite platform [3]. Among them, Terrain Observation with Progressive Scans (TOPS) mode has gained importance as it is the most efficient mode for mapping large areas [4,5,6]. The TOPS mode is considered a modification of the ScanSAR imaging mode [1]. With progressive scans in azimuth, it represents a primary acquisition mode (interferometric wide swath—IW) of the Sentinel-1 satellite due to its wide-area coverage capabilities. Unlike nadir-looking radar altimeters, the side-looking geometry provides range sensitivity and avoids ambiguous reflections since there is no spatial resolution directly under the platform. A radar onboard the satellite points its antenna directly to the Earth’s surface in a plane perpendicular to the flight direction. The position of the sensing instrument can be defined by a set of main geometric parameters that are depicted in Figure 1.

SAR satellites orbit the Earth in a sun-synchronous low Earth orbit (LEO) at altitudes approximately between 200 and 1000 km. Acquisitions are taken in both ascending (from south to north) and descending (from north to south) flight directions. The location of the image picture elements (pixels) can be defined in geographical coordinates or in slant range geometry. In the latter case, the azimuth and ground range directions are used as the axes of the image coordinate system. A curated list of synthetic aperture radar (SAR) software, libraries, and resources can be found in [7].

Remote sensing has become increasingly instrumental in the efficient and sustainable management of vineyards, offering a diverse array of applications ranging from crop growth monitoring [8] to soil moisture estimation [9,10]. Both satellites and unmanned aerial vehicles (UAVs) have been prominently featured in this context, providing comprehensive and detailed insights into vineyards across various temporal and spatial resolutions. Key applications of remote sensing in vineyards include vegetation cover mapping, irrigation monitoring, the early detection of diseases and pests, yield estimation, and even grape variety differentiation. These applications are vital for winegrowers as they enable more precise and efficient vineyard management, resulting in improved yields and reduced operational costs.

However, despite the evident benefits offered by conventional remote sensing, such as high-resolution optical imagery and high-precision UAV data, there are still challenges to be addressed. Limitations such as dependence on atmospheric conditions and the inability to conduct continuous monitoring under all weather conditions may constrain the effectiveness of these approaches. In this context, the use of SAR emerges as a promising solution to overcome these challenges and enhance vineyard monitoring capabilities. SAR offers several distinct advantages, including the ability to operate independently of weather and lighting conditions, as well as the capability to penetrate vegetation to provide detailed insights into soil and crop characteristics.

Therefore, exploring the potential advantages of SAR for vineyard monitoring applications is crucial for further advancements in precision viticulture. By leveraging the unique capabilities of SAR, it is possible to improve the quality and efficiency of vineyard monitoring, providing winegrowers with valuable insights to enhance their cultivation and management practices. In fact, the use of SAR in viticulture can already be found in several studies. The use of space-borne SAR in viticulture was reported in land cover classifications [11,12,13], soil moisture monitoring [14,15,16], and, particularly, in polarimetric applications of SAR data [17,18,19]. For example, in land cover classification, space-borne SAR can be used to differentiate between vineyard areas and other types of land cover such as forests or urban areas based on their radar backscattering characteristics. Additionally, in soil moisture monitoring, SAR data can provide valuable information about the moisture content of vineyard soils, which is crucial for irrigation management and predicting crop yield.

The repetitive capabilities of historical and contemporary SAR satellite missions (Table 1) and extended data capture enable the detection of temporal changes and trends, providing insight into the dynamics of viticulture regions. An SAR-based software for operational soil moisture mapping services in Italy was presented in Pulvirenti et al. [20], and the Copernicus Global Land Service has started providing surface soil moisture (SSM) of the soil’s topmost 5 cm over the European continent from Sentinel-1 sensors at 1 km × 1 km spatial resolution [21]. An overview and comparison of both methodologies are provided in van Hateren et al. [22]. Additional validation of SAR-derived surface soil moisture products is provided in Beale et al. [23] and in Balenzano et al. [24].

In this context, this article aims to provide a comprehensive overview of the use of SAR in viticulture and wine-growing areas. By synthesizing the existing literature and incorporating demonstrative insights, this article seeks to clarify both the limitations and advantages of current SAR applications in vineyard monitoring. Furthermore, this study also aims to offer a perspective on future trends, which appear highly promising given the number of missions currently in operation and the anticipated launch of new missions (Table 1). A systematic review using the Scopus database was conducted to provide a comprehensive background on SAR applications in viticulture, serving as a foundation for understanding the context and significance of our research team’s demonstrations and practical applications. The article aims to contribute to a deeper understanding of SAR’s role in viticulture and guide future research activities in this field.

Synthetic aperture radar (SAR) [1] is a coherent active microwave imaging method that has found widespread applications in various fields. It was initially developed for military reconnaissance and surveillance, offering high-resolution, all-weather, day-and-night imaging capabilities. SAR technology later expanded to civilian applications, including environmental monitoring, earth-resource mapping, and natural disaster management.

Synthetic aperture radar (SAR) is nowadays becoming a valuable tool for monitoring agricultural areas and supporting precision agriculture applications [25]. Several studies have reviewed the challenges of SAR monitoring in precision agriculture domains [26,27,28,29,30,31]; a comprehensive overview of the use of synthetic aperture radar (SAR) in the context of wetland and flood monitoring is presented in [32].

2. Remote Sensing Methods for Vineyard Monitoring

Remote sensing technologies like multispectral and hyperspectral imagery, thermal infrared imagery, and light detection and ranging (LiDAR) measurements have the potential to gather information on grapevine health, its canopy structure, water stress indicators, nutrient status, and grape quality. Various remote sensing platforms, including UAVs, aircraft, and satellite missions, have been used to characterize vineyard spatial variability and assess viticultural parameters [25,33]. These technologies are used to delineate vineyards from non-vineyard areas [34], discriminate grapevine varieties, and estimate grape quality and yield using spectral information [35,36]. Remote sensing platforms are capable of providing data in different temporal and spatial resolutions, enabling the study of viticulture regions. Vegetation indices [8] are used to predict grape phenolics and color at harvest, aligning observations with grapevine phenology [37]. The vineyard spatial variability [38] and its association with zonal harvesting and differences in wine sensory were demonstrated in Dorin et al. [39]. Optical satellite data, including multispectral imagery from satellites or UAVs, has been utilized for leaf area index (LAI) mapping and normalized difference vegetation index (NDVI) calculation for grape yield prediction [40,41]. Hyperspectral imaging has been employed to estimate yield and quality of grapes, monitor vineyard water status, and refine vegetation indices for precision agriculture, offering detailed spectral information for spatial variation determination in vineyards [42,43,44,45,46]. The integration of different remote sensing technologies, such as UAV-based hyperspectral, has shown promise in improving grapevine disease detection in vineyards, highlighting the potential for comprehensive monitoring and the early detection of stress factors [43]. Additionally, Global Navigation Satellite System Reflectometry (GNSS-R) has been applied to vineyard moisture monitoring, leveraging satellite data for assessing soil moisture levels in vineyards [47]. The synergistic use of optical and radar data mitigates the impact of meteorological conditions on remote sensing data and enhances agricultural product classification accuracy [48]. Despite the wide range of remote sensing applications in viticulture, Section 3 emphasizes the use of radar for: (i) grapevine monitoring and management; (ii) soil moisture estimation; (iii) land cover classification; (iv) InSAR for stability monitoring; and (v) airborne and ground-based SAR for vineyard monitoring.

3. Applications of Synthetic Aperture Radar in Viticulture

3.1. Grapevine Monitoring and Management

SAR imagery is advantageous for crop mapping as it is not affected by weather conditions and can provide timely information on crop type and conditions, helping agricultural planning and resource management.

The use of SAR data to estimate surface roughness parameters in vineyard soils is presented in Remond et al. [49]. The study area consists of vineyard plots in a catchment area that is prone to erosion. The researchers used the integral equation model (IEM) to analyze the behavior of the backscattering coefficient in terms of roughness, conducted field investigations, and compared the results with satellite data from the European Space Agency’s ERS-1 satellite. They found that the correlation between the observed and modeled values was 73%, indicating that SAR data can be used to describe agricultural surfaces with hydrological parameters needed for runoff modeling. Based on the results, the authors proposed a simple relationship between the backscattering coefficient and the roughness parameters, which could be used to describe different types of surfaces and their hydrological properties. They suggested that combining radar data from different angles or frequencies could improve the estimation of roughness parameters.

Burini et al. [50] analyzed ERS SAR images acquired in 1999 and 2001 to observe the backscattering coefficient and correlate it to wine production. The researchers developed a simple linear model to describe the relationship between the backscattering coefficient and various soil and vegetation variables, including fruit biomass and plant development. They found that the difference in backscattering between the vineyards and reference parcels could be linearly related to grape biomass. The study found a correlation between the change in backscattering and wine production per hectare, suggesting that SAR could be a potential tool for monitoring wine production. However, the limited revisit time of past SAR systems poses limitations to this application.

Schiavon et al. [51] explain that radar backscattering is sensitive to vegetation characteristics, and previous studies have shown its potential in monitoring crops like wheat or rice. However, monitoring grape biomass is challenging because the woody structure of vineyards is only slightly modified by the developing fruits. The presence of poles and metallic wires further complicates the measurements. Therefore, sophisticated processing techniques are needed to reduce interfering effects and accurately monitor vineyards. The authors conducted an experiment using polarimetric, very-high-resolution C- and L-band SAR data from an airborne SAR system. They compared the backscattering values before and after grape harvesting and observed an increase in backscattering during the summer months and a decrease in September/October, which was attributed to the development and subsequent harvest of grapes. The study also utilized a geographic information system (GIS) and ground truth data to validate the SAR measurements. The results suggested a sensitivity of L-band backscattering to grape biomass per unit area. However, the authors noted that further measurements and integration with optical/IR sensors are needed to better understand vineyard systems and develop grape retrieval methods.

Loussert et al. [52] highlight the variability in SAR and optical signals due to heterogeneities in vineyard organization and management. The researchers analyzed the effects of various vineyard organizations on SAR and optical data, including backscattering coefficients and polarimetric parameters. They also monitored the grapevine phenological cycle using measurements of green leaf area index (GLAI) and grapevine canopy height. The results showed that the inter-row spacing management had a strong impact on the temporal signatures, with variations in backscattering coefficients and polarimetric parameters. The researchers also found that the vegetation indexes increased with grapevine growth, depending on canopy width and soil management.

The potential of high-resolution X-band Cosmo-SkyMed imagery for grapevine vigor characterization was investigated in Roussel et al. [53]. The paper examines the correlation between SAR signal backscattering, grapevine variety, and rank orientation factors and their correlation with grapevine vigor using X-band Cosmo-SkyMed images and ground data. The study compared and analyzed the limitations and effects of various classification algorithms, regularization algorithms, speckle filtering, and image geometry. The results indicated that the relationship between X-band SAR signal amplitude and grapevine vigor may not be significant enough to build a prediction model. The paper suggests potential improvements, such as acquiring cross-polarized data, using steeper acquisition angles, and implementing interferometric conditions to enhance the accuracy of grapevine vigor classification using SAR imagery.

Beeri et al. [54] aimed to determine the crop coefficient (Kc) and LAI for grapevine irrigation using satellite imagery, and tested different methods for integrating SAR and optical sensors to improve temporal resolution and create seamless time-series of LAI and Kc estimations. The results showed that the Sentinel-1 SAR satellite achieved the best accuracy in estimating the crop parameters compared to Sentinel-2 and Landsat-8 optical satellites. The integration of all three sensors improved the overall accuracy of Kc and LAI estimations. The researchers concluded that Sentinel-1 can be a valid alternative to optical data for estimating Kc and LAI in vineyards, particularly in areas with cloud cover. The SAR imagery provided better accuracy in some cases, and its frequency of images ensured continuous monitoring of grapevine irrigation needs.

Davitt et al. [55] explored the use of NASA’s ECOSTRESS and the European Space Agency’s Sentinel-1A SAR satellite for crop monitoring, specifically focusing on grapevine growth and conditions in Sonoma County, California. The study assessed the complementary use of Sentinel-1A SAR and ECOSTRESS land surface temperature (LST) and evapotranspiration (ET) datasets to monitor vineyards. Correlations between SAR, LST, and ET measurements were identified, indicating that SAR can be utilized to identify grapevine growth and canopy development, offering valuable insights for vineyard management. The results showed that Sentinel-1A SAR backscatter measurements can provide indications of grapevine leafiness and moisture content, while ECOSTRESS LST and ET measurements can provide information on grapevine temperature and stress. The study also used radiometric modeling to understand the sensitivity of SAR backscatter to grapevine canopy structure and moisture conditions. The results suggested that SAR backscatter is primarily driven by grapevine canopy moisture content, with a slight sensitivity to grapevine leaf area index.

Fontanelli et al. [56] aimed to evaluate the potential of SAR imagery for early-season crop mapping, including vineyards, and to investigate the influence of various parameters on classification accuracy. Dual-polarized SAR data from the Cosmo-SkyMed X-band satellite mission combined with a three-dimensional (3D) convolutional neural network (CNN) classifier were used for early-season crop mapping. However, the study found that the classification accuracy of vineyards was poor, and further investigation is needed to understand the influence of vegetation phenology, structure, density, biomass, and turgor on the CNN classifier, along with the low producer accuracy marked by vineyards.

To demonstrate the applicability of SAR backscatter characteristics in vineyard monitoring throughout the year, temporal observations conducted at four vineyards in the Douro region of Portugal were used. For this, 28 ground range detected (GRD) SAR images acquired by the Sentinel-1A satellite in interferometric wide swath (IW) mode were used. These images were acquired during the period from January 2023 to December 2023 and are from the same acquisition track (52). All images are descending and feature VV and VH polarizations. Each of these images was processed by applying thermal noise removal, the precise orbit file, calibration, terrain correction, and conversion from linear to dB. At the end, ${\sigma }_{VV}^0$ and ${\sigma }_{VH}^0$ at four different points (vineyards) over 2023 were determined. The results showed that the ${\sigma }_{VV}^0$ and ${\sigma }_{VH}^0$ backscatter from Sentinel-1 responds distinctly to various aspects of vineyards, such as grapevine canopy development and precipitation occurrences, as depicted in Figure 2. The time series of ${\sigma }_{VV}^0$ tends to be higher than ${\sigma }_{VH}^0$, although both respond to precipitation events and grapevine growth stages. During higher precipitation periods in January, March, November, and December, there was an increase in the ${\sigma }_{VV}^0$ backscatter response, suggesting a correlation with radar sensitivity to precipitation and associated soil moisture conditions. On the other hand, ${\sigma }_{VH}^0$ backscatter demonstrated greater sensitivity to volume scattering from plants compared to ${\sigma }_{VV}^0$, showing an increasing response from April until the end of October. This period coincides with the active vegetative periods (flowering, maturation, harvesting) and subsequent leaf fall.

Vreugdenhil et al. [57] explored the potential of Sentinel-1’s ${\sigma }_{VV}^0$ and ${\sigma }_{VH}^0$ backscatter, along with the ${\sigma }_{VH}^0/{\sigma }_{VV}^0$ ratio (CR), to monitor crop conditions. This ratio, which combines co-polarized and cross-polarized backscatter, proved to be sensitive to volume scattering and exhibited an increase during vegetative growth. Figure 3 depicts the CR and ${\sigma }_{VH}^0$ in two separate vineyards, illustrating a clear peaking trend in July with a substantial increase observed from April to October. This confirms that the CR exhibits greater sensitivity to grapevine canopy growth, serving as a ratio that can be used effectively for this purpose. Thus, polarization is sensitive to different characteristics of the grapevine canopy, such as moisture and growth. Additionally, ${\sigma }_{VH}^0$ backscatter appears to be more responsive to changes in the grapevine canopy, whereas ${\sigma }_{VV}^0$ is more closely related to moisture. Consequently, it is feasible to implement and develop methodologies for vineyard monitoring and management.

3.2. Soil Moisture Estimation

In 2012, Ballester-Berman et al. [58] presented a polarimetric study on the use of C-band RADARSAT-2 data for monitoring vineyards in Spain. The researchers aimed to investigate the sensitivity of the scattering response of vines to different growth stages. They found that the scattering response from vines can be approximated as a two-component scenario, with a strong return from soil and a weaker depolarized return from grapes and leaves. Analysis of SAR data showed that the cross-polar backscatter intensity differed from the co-polar backscatter, indicating the presence of grape biomass. A water cloud model (WCM) was proposed to parameterize the difference in HV intensity as a function of changes in grapevine biomass, which can be used for vineyard monitoring and inversion purposes. Ballester-Berman et al. [59] discussed three different retrieval strategies for estimating soil moisture in vineyards: the two-component model, a classical three-component model improved by a physically constrained volume power, and a modification of the second model by including the IEM approach. The models and inversion algorithms were evaluated using RADARSAT-2 polarimetric data and ground-truth data on soil moisture, plant height, LAI, and phenology stage. The results showed that all three models matched the data well, with the IEM model for surface scattering at the C-band showing the best match.

Bazzi et al. [60] compared the accuracy of two soil moisture products derived from Sentinel-1/Sentinel-2 at the plot scale (S²MP) [15] and the Copernicus surface soil moisture (C-SSM) product in southern France. The S²MP product was obtained by coupling Sentinel-1 SAR data and Sentinel-2 optical data and provided soil moisture data over the agricultural areas at the plot scale with a six-day revisit time [15]. Copernicus Global Land Service distributes soil moisture estimations (C-SSM product) over the European continent at a 1 km spatial resolution using Sentinel-1 C-band data [21]. The results showed that both products have good agreement with in situ measurements, but the S²MP product is more accurate than the C-SSM product. The C-SSM tends to overestimate soil moisture values in forest and vineyard areas, while the S²MP product does not provide accurate estimations in areas with well-developed vegetation cover.

Lei et al. [61] discussed the integration of high-resolution thermal infrared (TIR) and Sentinel-1 SAR remote sensing data into a soil–vegetation–atmosphere transfer (SVAT) model to improve soil moisture estimates in vineyards. The study proposes a two-component polarimetric model for soil moisture estimation in vineyards using C-band radar data and underscores the potential of remote sensing data and data assimilation techniques for monitoring evapotranspiration and soil moisture in vineyards.

A methodology using soil moisture measurements, satellite images from Sentinel-1 and Sentinel-2, and terrain parameters to identify areas with low soil moisture content in a drip-irrigated vineyard during an agricultural drought is presented in Mendes et al. [62]. The study highlights the potential of using remote sensing data as a cost-effective solution for monitoring soil moisture content and predicting areas at risk of salinization in vineyards, especially considering the expected increase in droughts due to climate change. The authors also emphasized the importance of multi-sensor approaches, as different wavelengths have different sensitivities to field variables, and their integration can provide more reliable assessments. This research contributed to the understanding of the impacts of irrigation and drought on grape production and provides useful information for vineyard management in the face of changing climate conditions.

The use of Sentinel-1 and Sentinel-2 Earth observation data for predicting soil moisture content in vineyards using a cycle-consistent adversarial network (CycleGAN) is presented in Efremova et al. [63]. The researchers compared the performance of different machine learning (ML) algorithms, including linear regression, random forest, support vector regression, and neural networks. They found that the random forest algorithm performed the best, while the linear regression model showed significant overfitting. The study also evaluated the performance of the CycleGAN model for image translation between Sentinel-1 and Sentinel-2 data. The results showed that the model was able to extract some features from different domains, but the translation was not perfect.

The vineyard data assimilation (VIDA) system [64], which combines remote sensing data with a soil water balance model to improve irrigation decision-making in vineyards, showed promise in capturing daily variations in soil moisture. Accurate soil moisture monitoring is crucial for effective irrigation management in vineyards, with potential benefits including deficit irrigation. Remote sensing techniques like synthetic aperture radar and thermal-infrared sensing can provide high-resolution soil moisture monitoring. While the VIDA system has potential for improving irrigation decisions and support in vineyards, challenges include difficulties in correcting bias in irrigation inputs and accurately representing subsurface water-flow processes. Opportunities for improvement include considering DisALEXI ET observations and utilizing upcoming missions and Sentinel-2 data to enhance the accuracy and frequency of soil moisture and evapotranspiration retrievals.

The prediction of soil moisture samples based on Sentinel-1 and Sentinel-2 is presented in Figure 4. The model employed in this study relies on Sentinel-1 SAR backscatter data and water and vegetation indices derived from Sentinel-2 optical satellite imagery. The SAR data used consists of 60 series of IW level-1 GRD from three acquisition tracks (52, 125, and 147). SAR backscatter data were processed to create synthetic bands based on combinations of VV and VH polarizations from Sentinel-1. Synthetic bands were derived for Beta0, Gamma0, Sigma0 with terrain correction, and Gamma0 with terrain flattening. With optical data from Sentinel-2, the Normalized Difference Water Index (NDWI) [65,66], Normalized Difference Infrared Index (NDII) [67], and NDVI [68] were calculated. These values were extracted using the mean of a 3 × 3 window centered at each geographical coordinate. The geographical coordinates for point extraction were the same as for the sensors, placed in two vineyards. To construct the predictive model, a regression model was built using an artificial neural network (ANN) comprising six hidden layers. The main objective was to forecast soil moisture values relying solely on remote sensing data from Sentinel-1 and Sentinel-2. Following model training and validation, the predicted soil moisture values were compared against those measured by the on-site sensors. The evaluation yielded a coefficient of determination of 0.857 and a mean absolute percent error of 6.199.

3.3. Land Cover Classification

In Notarnicola et al. [69], multitemporal Sentinel-1 and Sentinel-2 imagery was used to detect phenological dynamics in different land cover types, including vineyards. The backscatter model simulations indicated a high sensitivity of the Sentinel-1 SAR sensor to plant and partially soil conditions, which can be used to monitor vineyards. Another exploitation of Sentinel-1 SAR data for land cover mapping, including vineyards alongside arable land and grassland, is presented in Orlíková et al. [70].

A land cover map using 76 polarimetric parameters extracted from a quad-polarized ALOS PALSAR image and an object-oriented classification method was produced in Moine et al. [71]. Nine land cover classes were mapped, with a success rate of 69%. The well-identified classes are forest, urban areas, and water. Other classes, such as vineyards, meadows, and cultivated fields, were more difficult to classify due to their similar polarimetric signatures. Despite this, the method showed promising results for land cover mapping based on polarimetric SAR images.

3.4. InSAR for Stability Monitoring

The application of SAR interferometry (InSAR) has been pivotal in monitoring slope instability and ground displacements, demonstrating its relevance in geohazard assessments [2,72,73,74,75]. SAR technology is likewise instrumental in providing information on vineyard displacements, which is essential for ensuring vineyard safety and stability.

InSAR applications that are specifically for vineyard monitoring are rarely found in the literature. This is because interferometric applications require coherent or persistent scattering targets on the ground (man-made objects, exposed rocks, non-humid soils), which would allow for the reconstruction of interferometric phase measurements in subsequent satellite acquisitions. Vegetation or grapevine fields are usually non-coherent areas and thus do not correspond to objects of interest for InSAR analysis. The radar backscattering response from vegetation is too noisy for interferometric use due to vegetation growth and leaf changes, and is thus mitigated as temporal decorrelation in InSAR analysis.

Nevertheless, SAR has been used for deformation monitoring, and several modifications to the standard persistent scatterer interferometry (PSI) technique [72,73] for surface deformation monitoring [76,77,78,79,80,81,82,83,84], including techniques examining distributed and partially coherent radar scattering targets [85,86,87,88], have been introduced into the field, indicating InSAR’s capability to detect subtle changes in the vineyard terrain or terrain with shallow vegetation. Moreover, SAR interferometry could be an effective monitoring technique for geologists and regulators overseeing underground activities such as mining [89], demonstrating its potential for subsidence monitoring, which could also be relevant for vineyard management. An example of a regional-scale InSAR deformation map with mean line-of-sight (LOS) velocities from PSI processing of Sentinel-1A/B’s ascending track No. 147 is shown in Figure 5.

The area of interest covers approximately 100 by 60 km (≈6000 km²) (Figure 5). A total of 420 SAR images spanning 9 years of observations (27 October 2014–21 November 2023) were analyzed by persistent scatterer interferometry (PSI) technique [72,73] implemented in SARproZ software [90]. All available single-look complex (SLC) images were co-registered on the sampling grid of a primary acquisition (28 June 2021) and were selected by optimizing the distribution of perpendicular and temporal baselines and assessing weather conditions [91]. The data were analyzed on a pixel-by-pixel basis to identify a sparse grid of points, typically corresponding to man-made structures (e.g., retaining vineyard walls, roads, buildings, etc.). Precise orbit ephemerides (POE) and the SRTM-3 digital elevation model were used for the processing. A threshold of 0.65 on the amplitude stability index (ASI) was used to create a network of pre-selected points for estimating preliminary parameters and the atmospheric phase screen (APS). After APS compensation, the residual height, displacement velocity, and displacement time series were estimated with respect to a reference point selected in a scatterer with high amplitude stability (ASI > 0.9), located at 41.17148408°N latitude, −7.35213141°W longitude. A displacement map with estimated mean line-of-sight velocities is shown in Figure 5. The impact of ground motion on vineyard areas and the surrounding infrastructure can be monitored at approximately 1.2 million persistent scattering points over the entire Douro wine region. Examples from PSI deformation monitoring are shown for the infrastructure (Figure 6a,c) and vineyard areas (Figure 6b,d).

Point scattering targets in Figure 5 and Figure 6 are visualized with colors proportional to line-of-sight (LOS) velocities (mm/year). The amount of displacement is represented by a colorbar shown on the right of Figure 5, with displacement rates ranging from −5 (red) to +5 (blue) millimeters per year. Positive values (blue color) of LOS velocities correspond to displacement towards the satellite, i.e., uplift. Conversely, red areas indicate subsiding regions or motion away from the satellite.

3.5. Airborne and Ground-Based SAR for Vineyard Monitoring

The airborne SAR platforms are also used in monitoring vineyards. Polarimetric data provide a more detailed description of the environment compared to single-channel radar systems. The potential of polarimetric data acquired by the airborne RAMSES SAR in bands L and P for characterizing vineyards in southwest France was investigated by Baghdadi et al. [92]. The results showed that the polarimetric SAR data in L- and P-bands can effectively discriminate between land cover classes but are not suitable for classifying vineyards based on row direction or age.

The potential of airborne radar remote sensing to characterize vineyards was also demonstrated in Burini et al. [93,94]. The study discussed the use of airborne polarimetric SAR and QuickBird optical data for vineyard monitoring and biophysical parameter retrieval. The study was conducted in the Frascati area near Rome, Italy, as part of the BACCHUS project, which aimed to establish a high-quality geographic information system for vineyards. The DLR ESAR synthetic aperture radar system flew over the study area in October 2005 at a high spatial resolution in the C-band and L-band polarimetric modes. The multi-frequency configuration, which used both L-band polarimetric SAR and C-band dual-polarization SAR data, was found to be the most effective in detecting vineyards. The researchers employed a supervised neural network algorithm to classify the vineyards and olive groves in the study area using radar and optical data. The accuracy of the classification was evaluated using confusion matrices, and the results were compared between the different datasets. The analysis showed that the combination of radar and optical data improved the overall accuracy of the classification, particularly in discriminating between different land cover classes like vineyards, olive groves, grass/bare soil, and urban areas. Furthermore, the study explored the potential of SAR data in detecting variations in vineyard vigor. The SAR characteristics, such as backscattering, were correlated with ground truth measurements of vineyard vigor to assess the capability of high-resolution radar in detecting in-field variations.

Regarding ground-based SAR, a system for detecting grape bunches was presented in Eccleston et al. [95]. This study describes an SAR operating between 2 and 6 GHz that was used for detecting grape bunches and predicting grape yield in a vineyard. The 3D images obtained from the SAR system showed that it was able to penetrate the leaves and accurately detect the grape bunches even when they were occluded by leaves. In Parr et al. [96], experiments using a highly directional ultrasound array were conducted, and a computer-controlled gantry was used to scan grape bunches behind grapevine foliage. The authors found that low-frequency ultrasound signals can propagate through foliage and reflect off the grapes behind. This promising approach could provide a non-destructive remote-sensing method for accurately estimating grape yield in vineyards. In comparison to ultrasound methods, the authors reported that SAR can be used as an alternate option, but it was found to be less cost-effective.

4. Future Perspectives of SAR in Viticulture

The potential of satellite SAR for viticulture and vineyard monitoring is set to expand dramatically with the upcoming launch of several new missions. These include ESA Biomass, NISAR, ROSE-L, HARMONY, TanDEM-L, ALOS-4, and Sentinel-1 C/D (Table 1). These missions can significantly improve SAR capabilities in agricultural and viticulture monitoring by offering advancements in resolutions, polarizations, revisit periods, and wavelengths. The combined use of different wavelengths in SAR applications, including viticulture, offers significant benefits. Longer wavelengths (like L-band or P-band) penetrate deeper into the vegetation canopy than shorter wavelengths (like X- or C-band). While shorter wavelengths are more sensitive to changes in smaller vegetation elements and moisture content, longer wavelengths can provide information on both the surface and subsurface features of the vineyard. Combining different wavelengths enhances soil moisture monitoring accuracy, improves resolution and coverage, mitigates temporal and weather sensitivity, and improves feature differentiation, enhancing land cover classification, vineyard mapping, and management practices overall.

The NISAR mission, set to launch in the near future, has significant potential for monitoring vineyards and viticulture. The quad-polarimetric images generated by NISAR in L-band and S-band frequencies present an opportunity for accurate mapping and monitoring of above-ground biomass on a global scale, which is essential for assessing vineyard health and productivity [97]. NISAR will provide users with polarimetric and interferometric data, enabling large-scale land cover monitoring [98,99]. The mission’s capability to acquire fully polarimetric L-band SAR data can be instrumental in monitoring vineyards due to its ability to capture detailed information about vegetation structure and health [100]. Additionally, the NISAR mission aims to retrieve biomass from dry forests and woodlands, showcasing its relevance for biomass estimation, which is crucial in vineyard monitoring [101]. The NISAR mission’s dual-band radar instrument design and global persistent SAR sampling capabilities highlight its potential for comprehensive and continuous monitoring of land surfaces, including vineyards [102,103].

Biomass, which includes the P-band wavelength to penetrate deep into the canopy, represents a significant milestone in forest and biomass observation [104,105,106]. The combined use of data from different SAR sensors, such as Biomass or ROSE-L, which employs L-band SAR specifically designed to monitor larger vegetation structures and aims to fill the observation gaps of the existing group of satellites, serve as prime examples of the unique utility of these missions within the field of viticulture. The SARSense campaign [107] has demonstrated the potential of ROSE-L for applications such as soil moisture monitoring, irrigation management, crop type discrimination, and precision farming, which are all critical for the effective management of vineyards.

The European Space Agency (ESA) and European Commission (EC) have proposed to launch the third (Sentinel-1C) and fourth (Sentinel-1D) units of the Sentinel-1 constellation. Similarly, Sentinel-1-Next Generation (NG) C-Band SAR is focused on enhancing data continuity beyond 2030 [108]. In addition, the ESA will launch dual SAR satellites, known as Harmony, which will synergize with Sentinel-1. Harmony’s data could complement information obtained from other SAR missions focused on agriculture, providing a more holistic view of environmental conditions and improved fine-resolution climate models to assess micro-climatic variations within vineyards.

The advancements from ALOS-2 to ALOS-4 [109], with Japan’s pioneering L-band SARs and cutting-edge capability to observe 200 km swaths in high-resolution stripmap (SM) mode, also indicate significant improvements in global vegetation monitoring.

Moreover, due to the emergence of small satellite SAR missions led by companies such as ICEYE, Capella Space, Synspective, Umbra Space, and iQPS, the overall SAR application landscape is poised to bring substantial transformations in SAR applications within the fields of precision agriculture and viticulture. The implementation of these new satellites enhances both the level of detail and the frequency at which data are collected.

At the data processing level, Ferro and Catania [25] examine the role of artificial intelligence (AI) in various applications in grape production and explain the benefits of precision viticulture in terms of stress identification, vineyard spatial variability, and environmental impact. Their study also highlights the use of ML and deep learning algorithms for tasks such as yield prediction, disease detection, and grape quality assessment. The paper emphasizes the importance of image processing, computer vision techniques, and the potential of Internet of Things (IoT) technologies in vineyard monitoring, as well as outlining future scenarios for viticulture, including the development of rapid and detailed real-time results and the interconnection of different technologies to support sustainable production patterns. Additionally, it discusses the role of unmanned ground vehicles (UGVs) for site-specific agronomic management and the potential of IoT technologies to relay comprehensive information in real time for farmers or robotic platforms.

By analyzing vast datasets from various sources, such as satellite images, ground sensors, and drones, AI algorithms can provide insights into yield prediction, disease detection, and stress identification with high accuracy. Deep learning models, trained on historical and real-time data, will be able to predict future vineyard conditions, enabling preemptive actions to optimize grape production and quality. IoT devices and wireless sensor networks deployed across vineyards can stream data continuously, providing a real-time snapshot of various parameters like soil moisture, temperature, grapevine health, and more. An overview of sensors used in vineyard monitoring [25] is provided in

Table 2. Main sensors and platforms for vineyard monitoring.

Sensor	Description	Measured Characteristics
Multispectral sensors	These are the most commonly used sensors in vineyard monitoring. They typically record radiation reflected by grapevines in a small number of broad bands, between 2 and 8, often used to detect stress conditions. Commonly include five sensors for blue, green, red, red-edge (700–740 nm), and near-infrared (780 nm) wavelengths.	Grapevine health, stress levels, photosynthetic activity, chlorophyll concentration, and overall plant vigor.
RGB cameras	Used for object detection surveys in vineyards, identifying grapevine canopy, shoots, or bunches. Operate within the wavelengths of blue, green, and red.	Canopy structure, shoot density, bunch presence, visual health indicators, and plant growth patterns.
Hyperspectral sensors	Used to characterize water status, grape quality, and the early identification of pathogens and diseases. Collects data across a wide range of wavelengths (400–2500 nm) with high spectral resolution.	Water status, grape quality parameters (e.g., soluble solids, anthocyanins), early detection of pathogens and diseases.
Thermal infrared sensors	Mainly used to investigate grapevine water stress, particularly for irrigation purposes. Measure spectral ranges of 7500 to 14,000 nm and can detect thermal variations.	Grapevine water stress, canopy temperature, irrigation requirements, and heat stress indicators.
LiDAR sensors	Used to estimate biophysical parameters of the grapevine canopy, such as height, width, and density. Include technologies using laser signals for high-precision measurements.	Canopy height, width, density, volume, precise terrain elevation, and spatial variability within the vineyard.
Optical remote sensing satellites	Various satellites like RapidEye, Landsat 8, IKONOS, Quickbird, MODIS, ASTER, WorldView-3, Pléiades, Sentinel-2, and many others are used. Provide data for monitoring vineyard variability, evapotranspirative processes, soil moisture content, and vineyard water stress prediction.	Soil moisture levels, vineyard layout, structural integrity of grapevines, and biomass distribution.
Unmanned aerial vehicles	Offer high-resolution data and are efficient and flexible at acquiring multispectral data, thermal infrared images, and RGB images for photogrammetric processing.	Canopy vigor, water status, grapevine growth patterns, spatial variability, and comprehensive monitoring of vineyard health.
Proximal sensors	Include optical contact sensors and soil moisture sensors used to measure soil properties at close range for assessing the nutritional and physiological states of vines or irrigation. Examples are GreenSeekerTM, Crop CircleTM, OptRXTM, and handheld thermal cameras for measuring parameters like chlorophyll content, nitrogen status, and crop water stress index (CWSI).	Leaf chlorophyll content, nitrogen levels, disease symptoms, photosynthetic efficiency, soil moisture, and plant stress responses.
Synthetic aperture radar (SAR)	Uses radar to capture high-resolution images of the Earth’s surface, including roughness, structure, and dielectric properties. Often used in satellites for all-weather, day-and-night monitoring.	Provides data for monitoring vineyard variability, soil moisture content, vineyard structure, and stress detection.

.

Integrating various sensors, such as multispectral, hyperspectral, thermal, LiDAR, and SAR, along with UAVs and proximal sensors, can significantly enhance real-time decision support monitoring systems for vineyards. Multispectral and hyperspectral sensors provide detailed insights into vegetation health, stress levels, and chlorophyll content, while thermal sensors detect grapevine water stress and temperature variations. LiDAR sensors offer precise measurements of canopy structure and terrain elevation, and SAR enables all-weather, day-and-night monitoring of soil moisture and vineyard structure. UAVs provide high-resolution imagery and 3D mapping, complementing data from optical remote sensing satellites. Proximal sensors, including optical contact and soil moisture sensors, deliver ground-level measurements of grapevines’ nutritional and physiological states. By integrating these diverse data sources into a centralized system, vineyard managers can monitor and analyze critical parameters in real time, enabling proactive management decisions to optimize grapevine health, irrigation practices, and overall vineyard productivity.

The integration of various technologies holds great potential for enhancing vineyard management practices. By combining different data sources, we can achieve a more comprehensive understanding of vineyard conditions and improve decision-making processes [110,111]. Data from different sources should be synthesized in a unified platform, providing a holistic view of the vineyard’s health. This integration will facilitate more sophisticated analytics, predictive modeling, and decision-making processes, all in real time. Vineyard automation tasks can be extended beyond data collection. Robotic platforms and UGVs can carry out a range of tasks from pruning grapevines to grape harvesting. These robots, guided by AI and real-time data, can perform tasks with precision, reducing labor costs and minimizing human error.

The interconnection of these advanced technologies will support sustainable viticulture practices. Precise monitoring and management will lead to the optimized use of resources, reduced waste, and minimized environmental impact, contributing to the sustainability of vineyard operations and, ultimately, better decision support.

5. Conclusions

Viticulture plays a vital role in the local economies of various countries. Though only 2.0% of the EU Utilized Agricultural Area (UAA) is covered by vineyards, it accounts for a significant percentage (45%) of the world’s total wine-growing areas [112]. In 2020, vines were grown on 3.2 million hectares in the EU. Spain, France, and Italy together accounted for three-quarters (74.9%) of the area under vines in the EU and about two-fifths (38.7%) of vineyard holdings in 2020 [112].

Better grape productivity requires regular vineyard monitoring. Satellite remote sensing can complement traditional vineyard monitoring methods. SAR satellite sensors provide all-weather, wide-area, day-and-night, high-resolution imagery, surpassing some of the limitations of other remote sensing approaches. The recent SAR satellite technologies have the potential to continuously monitor vineyards by understanding their spatial changes and development. Irrigation monitoring at regular intervals is also an important consideration when making irrigation practice decisions. Throughout the season, the vineyard canopies undergo a variety of agro-technical activities that influence the plants’ water intake. Several approaches reported in this study showed that Sentinel-1 imagery can be used to study crops’ relative water demand, crop coefficients, and leaf evapotranspiration, as well as LAI. The soil water content has a significant impact on the yield, growth, and quality of the wine. The effective retrieval of soil moisture can be facilitated by high-resolution SAR sensors.

Climate change is negatively affecting vineyards globally [113]. The primary issues that vineyards face are early ripening, increased sugar content, water constraints, and drought. Several studies have shown that vineyards can act as effective carbon sinks and storage systems [114,115]. This emphasizes the importance that permanent cropping systems can play in reducing greenhouse gas emissions (CO2). This role may be critical, particularly in regions with significant human demands, such as peri-urban agricultural spaces, where the preservation of grapevine ecosystem services may represent an attempt to improve environmental quality and, hence, viticultural viability. Brunori et al. [114] also indicate how vineyard management approaches, particularly soil management, may influence the grapevine’s carbon sink functions. By examining grapevine ecosystem services, particularly above-ground carbon storage and accumulation, Williams et al. [116] found that vineyard woody biomass carbon concentrations rise over time.

The above-ground biomass of grapevine plants can potentially be estimated through the upcoming Biomass mission, the first satellite to carry a full polarimetric P-band SAR. The mission primarily focuses on measuring forest biomass, but the P-band radar waves, with their deep canopy penetration, exhibit varying interactions with different agricultural structures. The incorporation of satellite SAR technology represents a prospective scenario in which vineyard monitoring becomes more accurate, reliant on data analysis, and supportive of sustainable viticulture methodologies. The capabilities of next-generation L-band SAR satellites (such as NISAR, ROSE-L, TanDEM-L, and ALOS-4), availability of high-resolution imagery (Cosmo-SkyMed, TerraSAR-X), continuity of C-band monitoring scenarios (Sentinel-1 C/D, Harmony), and small-sat monitoring constellations are set to revolutionize viticulture by providing more detailed and frequent data, facilitating better management and monitoring of vineyards.

The IoT, AI, hardware, software, and service linkages are revolutionizing precision viticulture. These technologies enhance real-time monitoring, data processing, and decision-making, increasing productivity and grape quality. Vineyard monitoring will benefit from modern sensors like satellite SAR missions, drones, GPS, and data analytics for yield monitoring and irrigation control. The viticulture industry is adopting a holistic approach, using AI and ML for data analysis and IoT to construct interconnected vineyard monitoring systems, all while focusing on sustainable wine-growing practices. Despite high initial expenditures and the requirement of technology literacy among growers, educational programs and the development of cost-effective, user-friendly technologies are crucial for the growth of the viticulture sector.