Olive tree drought stress: A systematic review

AI

Artificial Intelligence

ANN

Artificial Neural Network

CWSI

Crop Water Stress Index

DNN

Deep Neural Network

FAO

Food and Agriculture Organization

GRVI

Green Red Vegetation Index

Gs

Stomatal Conductance

GHG

Greenhouse Gas

GNDVI

Green Normalized Difference Vegetation

HS

Heat Stress

LC

Leaf Conductance

LPCP

Leaf Patch Clamp Pressure

LWP

Leaf Water Potential

NDVI

Normalized Difference Vegetation Index

PRI

Photochemical Reflectance Index

RCM

Regional Climate Model

RDI

Regulated Deficit Irrigation

Rs

Stomatal Resistance

RWC

Relative Water Content

SP

Stem Potential

SPI

Standard Precipitation Index

SHD

Super High Density

SWP

Stem Water Potential

SR

Stomatal Resistance

TDV

Trunk Diameter Variation

UVB

Ultraviolet B

The olive tree (Olea europaea) is an ancient tree species that has been found for many years in the Mediterranean region (Ninot et al. 2018; Poveda & Baptista 2021). The Mediterranean region accommodates at least 90% of the world's olive trees (FAOSAT 2018); however, it is currently threatened by climate change impacts, especially drought stress. Climate change is a global issue that is affecting many geographical areas, particularly the Mediterranean region (Fouial et al. 2016; Fraga et al. 2020b; Cortignani et al. 2021), and it significantly reduces agricultural activities due to abiotic stresses such as drought and salinity. Drought stress is a complex form of stress that impacts various aspects of plants, including their morphological, physiological, biochemical, and molecular characteristics (Lisar & Hamideh 2016; Mustapha & Mhamed 2024). These are generally negative transformations that lead to a reduction in the crop yield and even crop death in the worst scenario (Zia et al. 2021), resulting in extreme hunger, poverty, and acute malnutrition, respectively.

Therefore, it is essential to employ every available tool for monitoring, detecting, and managing olive survival under climate change, particularly drought stress. Several machine learning algorithms have been proposed to detect drought stress in plants, such as support vector machine (Ramos-Giraldo et al. 2020; Dao et al. 2021) and deep learning techniques (Lee et al. 2018; An et al. 2019). Climate change is associated with erratic rainfall, increased evapotranspiration, and reduced groundwater globally, particularly in the Mediterranean arid region, which produces more than 90% of the world's olive oil and table olives (FAOSAT 2018). This poses a potential threat to the region's agricultural sector and other consumers of olive-derived products. A study by Fernández (2014) revealed that drought stress has a negative impact on the physiology, growth, and development of olive trees. The olive tree is one of the Mediterranean countries' main economic crops, providing income and food to people throughout the region, especially in rural areas. Moreover, the olive tree is generally known to be a drought-resistant tree. However, according to Brito et al. (2019a, b), the future climate of the Mediterranean will affect the physiology and yield of olive trees. Numerous reviews have been conducted with respect to crop drought stress (Mechri et al. 2020; Zhou et al. 2021); however, only a few studies have been dedicated to reviewing olive tree drought stress (Brito et al. 2019a, b; Fraga et al. 2020b; Tramblay et al. 2020). The majority of the reported review articles are conventional literature reviews, which lack in-depth analysis and synthesis of the investigated research articles and do not follow a systematic literature review (SLR) approach. Thus, it is imperative to do a thorough and systematic literature analysis on this important topic. It is in view of this that we propose an SLR on olive tree drought stress based on the existing literature, surveys, and original research available to shed light on this dilemma. In addition, to the best of the researchers' knowledge, this is the first work that holistically reviewed olive tree drought stress following a systematic approach, which also contributes to the novelty of the work.

The main contributions of this study are as follows:

• This study provides a comprehensive evaluation following a systematic approach to investigating olive tree drought stress monitoring, methodologies, experiments, datasets, challenges, and potential future works.

• The study also revealed that the diversity of experimental approaches analyzed highlights the complexity of studying drought stress in olive trees.

• The findings discovered that although Turkey, Morocco, and Egypt are significant olive producers, they are underrepresented in olive tree drought stress research.

• To the best of the researchers' knowledge, this is the first work that proposes a holistic artificial intelligence (AI)-based framework for olive tree drought stress monitoring and mitigation under climate change. This resource sets the foundation for future studies on olive tree drought stress in the context of climate and AI.

The remaining part of this article is organized as follows: Section 2 discusses the related works, and the methodology is discussed in Section 3. Section 4 includes the result distribution and analysis.

Section 5 discusses the limitations and future works, while Section 6 presents the conclusion.

There are a number of related review articles on olive tree drought stress reported in the literature. For example, Fernández (2014) focuses on the adaptation of olive trees to abiotic stresses such as drought, salinity, and extreme temperatures. The review highlights the physiological and molecular mechanisms that enable olive trees to adapt to these stressors, including changes in gene expression, osmotic adjustment, and antioxidant defense systems. The authors explore the potential use of biotechnological approaches, such as gene editing and genetic engineering, to enhance olive tree adaptation to abiotic stress. He revealed that this could lead to the development of new olive varieties that are more resilient and productive in challenging environments, which could have significant implications for the olive oil industry. Moriondo et al. (2015) reviewed models that have been made to simulate olive tree and grapevine yields. The authors focused on the empirical models, which use the statistical relationship between climate and yield, and the process-based models, in which crop behavior is defined by a set of relationships that describe the main plant processes. Brito et al. (2019b) reported a review of the global status and significance of the olive trees' ecological system, as well as the effect of environmental abiotic stresses on olive cultivation. The authors examined and evaluated the negative effects of drought, the most crucial stressor to agriculture in the Mediterranean basin, and they also explored the main reactions of olive trees. Similarly, Fraga et al. (2020b) discussed olive adaptation strategies under climate change and its impacts, and they revealed that a single adaptation strategy cannot sustain the Mediterranean olives due to a number of issues. The approaches used by the authors for adaptation include short-term and long-term adaption strategies.

Zhou et al. (2021) present the current practices of using infrared thermal imagery to assess crop drought stress (also called water stress). Their study highlights the potential use of deep learning approaches for assessing drought stress. However, the authors did not give detailed information on several aspects related to crop drought stress assessment using thermal imaging and its limitations. Also, Crespo et al. (2024) only considered remote sensing approaches in their olive tree drought stress study.

However, these review papers did not present any experimental data or new findings; rather, they provide a summary of the existing knowledge on the subject. In addition, neither specific recommendations nor detailed information on the implementation of agronomic strategies to mitigate the effects of drought on olive trees are provided. Moreover, none of the existing review papers on this topic follow the SLR approach, and the tools, techniques, methods, and experiments used in the detection and monitoring of olive tree drought stress were only partly considered. As a result, this study is being conducted to fill the gaps in the reviewed papers and provide the current state of the art in olive tree drought stress studies within the framework of climate change. Table 1 summarizes the comparison of similar review papers found in the literature.

This study follows an SLR to analyze and synthesize olive tree drought stress. The literature review particularly sticks to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses standards (Liberati et al. 2009; Abdulganiyu et al. 2023) in addition to the proposed novel framework. Table 2 presents the distribution of journals and conferences, while Table 3 shows the different publishers of the selected papers.

The following planning protocol was used for the review: a review of the background definition, an assessment of the need for detailed SLR, the development of the study framework, research questions, a searching strategy, quality assessment (QA), data extraction, study selection criteria, and data analyses.

• Q1: What are the methods or techniques used to detect olive trees' drought stress?

• Q2: What are the data or datasets used to detect olive trees' drought stress?

• Q3: What are the reported olive tree drought stress experiments?

• Q4: What are the climate change impacts and adaptation strategies of olive trees?

• Q5: What are the challenges facing olive tree drought stress studies?

The proposed study seeks to answer the five research questions raised earlier. Previous research addressed some of these questions partly, but not all of them, despite the urgency.

This section describes how the literature search was carried out and what was discovered to answer the research questions proposed in Section 3.2. The databases used in the literature search are Science Direct, MDPI, IEEE Xplore, Springer Nature, Taylor and Francis, etc. The keywords used were olive tree drought stress, olive tree adaptation strategy, climate change, and olive tree drought detection methods. By using specific keywords developed in response to the five research questions, an automated search strategy was used to investigate and obtain relevant research articles. The key phrases used to construct the search terms had been selected based on terms commonly used in the literature and terms pertinent to this research. The phrases used in the search of primary research are as follows: (‘olive tree’ OR ‘olive orchard’) AND (‘Adaptation’ OR ‘Climate change’ OR ‘drought stress’ OR ‘abiotic stress’ OR ‘drought monitoring’ OR ‘drought detection’). The keywords and phrases used in searching for the literature reviews and surveys are as follows: (‘olive tree drought stress review’, ‘olive tree drought stress overview’, and ‘olive tree drought stress survey’).

During the primary selection process, duplicate copies were removed and the remaining records were evaluated based on their title, abstract, and subject focus. This was done to exclude sources such as lecture notes, reports, and books that were not published in journals or conference proceedings. As a result, the range of eligible works for publication in credible journals and proceedings was limited. Eligibility criteria were developed based on research questions and were used to filter the primary records and create secondary records that only included relevant studies for data analysis.

• Research articles related to olive trees.

• Research articles related to drought stress.

• Published articles in peer-reviewed journals.

• Research articles published before the year 2008.

• Research articles not related to olive tree.

• Research articles published in predatory journals.

• Research articles irrelevant to drought stress.

The secondary articles chosen were subjected to a QA to demonstrate the importance of the evaluated papers to the contribution of this work. To assess the quality of various investigations, it is critical to follow a few steps and use a quality evaluation instrument. The best articles for the study were picked by assessment by utilizing a series of five quality assurance questions, which are as follows:

• QA1 Does the topic of the paper relate to olive tree drought stress?

• QA2 Does the study provide a straightforward explanation of the background issue it addresses?

• QA3 Is the methodology of the study explained and interpreted appropriately?

• QA4 Does the research paper report the results of the experiments performed?

• QA5 Does the paper suggest further future research?

Each of the aforementioned five questions is assigned a score of one, so papers that meet all five criteria will receive a total score of five. In rating each paper based on the five criteria, two factors must be fulfilled to classify a paper to be of high quality and included in the analysis:

• A paper has to get at least three total points.

• A paper must satisfy criteria QA1, QA3, and QA4.

Each of the selected studies was examined for pertinent data and information. The extracted data included the paper's methods or techniques, experiment duration, datasets, territorial scope, as well as the study's findings.

Insightful findings regarding olive tree drought stress have been reported in numerous research publications. These insightful papers examine drought monitoring approaches or methods, experiments, climate change impacts, and adaptation measures. This research was conducted across several geographical areas with different research settings. Let us explore these fascinating works to gain a deep understanding of olive tree drought stress using Table 4.

The findings revealed that Turkey, Morocco, and Egypt are significant producers of olives but are not adequately represented in the field of olive tree drought stress studies.

Table 5 focuses on the distribution of articles based on the techniques used, along with their associated advantages and disadvantages.

In response to research questions 1 and 2 (Q1 and Q2), based on the studies analyzed, various approaches/techniques (Gs, leaf water potential (LWP), stem water potential (SWP), leaf conductance (LC), artificial neural network (ANN), deep neural network (DNN), normalized difference vegetation index (NDVI), photochemical reflectance index (PRI), logistic regression, and phytochemicals) were separately employed by a number of studies for olive tree drought stress. Other studies used a hybrid of more than one technique (NDVI + green red vegetation index (GRVI), SWP + LC, LWP + growth + yield, LWP + Rs + Gs, pythochemicals + UVB, SWP + LC + stem potential (SP) + trunk diameter variation (TDV)) to make the drought detection more reliable. Researchers select the techniques based on their research objectives, the sensitivity and precision of the techniques, the need for nondestructive or destructive measurements, the scope of the observations, available resources and expertise, and the desire for a comprehensive understanding. Recent olive tree drought stress studies have used deep learning techniques due to their capacity to manage large and complex datasets, automatically extract relevant features, and accomplish high predictive accuracy. While there is no single best method or technique for detecting drought stress that can be applied in all circumstances, stomatal conductance correlated best with crop water stress index (CWSI), surpassing leaf and SWP techniques. Each technique has advantages and disadvantages, as shown in Table 5, and the choice should be based on the requirements, objectives, and limitations of a study.

In the context of drought detection and analysis of datasets, researchers have employed different types of data, the majority of which are in situ data from olive trees, climate data, and remote sensing data as shown in Table 4. In situ data refer to measurements recorded directly from fields, including soil moisture content, LWP, and tree growth rates. This type of data provides highly accurate and detailed information specific to the monitored trees, enabling researchers to assess the direct impact of drought on olive tree health. The disadvantage of relying exclusively on in situ data is that they may have limited spatial coverage, making it difficult to capture drought conditions in larger regions.

Climate data, on the other hand, involve the collection of weather-related variables, such as temperature, humidity, and precipitation, from weather stations. Climate data offer broader coverage, capturing drought conditions over larger areas. It enables researchers to analyze long-term trends and patterns, thereby facilitating a better understanding of drought events. However, climate data provide generalized information and may not accurately reflect drought conditions in specific locations.

Remote sensing data involves the use of satellite or airborne sensors to measure indicators related to vegetation health and water availability, such as vegetation indices, thermal imagery, or soil moisture content. Remote sensing provides wide spatial coverage and can provide valuable insights into the overall vegetation health and identify areas prone to drought stress. However, its spatial resolution may be limited and it may not capture specific tree-level details. The remote sensing imagery datasets such as Planet, Sentinel 1, Sentinel 2, and Sentinel 3 provide valuable geospatial data, but they lack preexisting labels, which makes the process of labeling them a time-consuming and labor-intensive task. Manual inspection and interpretation of each data point are required to assign accurate labels, demanding domain expertise and extensive effort.

In summary, researchers use in situ data for accurate and detailed information about specific olive trees but with limited spatial coverage. On the one hand, climate data provide broader coverage and lack location-specific accuracy, while on the other hand, remote sensing data offer wide spatial coverage and quick monitoring and lack detailed tree-level information. Thus, combining different satellite data and sensors can provide a more comprehensive understanding of olive trees' drought conditions.

In response to research Q3, the result of the analysis and the summary in Table 4 show that there is considerable variability in the designs and methodologies of olive tree drought stress experiments reported in the literature. It is revealed that while the olive tree is generally known to be drought resistant, the effects of drought vary depending on when the drought began. The table olive cultivators should irrigate more than the olive oil cultivators because the latter need less water to maximize yield. The drought has no effects on yields if it occurs after July because fruit set has already occurred in the Mediterranean; this may be different in other parts of the world. Here are the primary considerations while conducting the drought experiments:

• Duration of experiments: The duration of experiments can range from a few days to several weeks, and in some cases from a few months to several years.

• Irrigation methods: The majority of researchers are using drip irrigation because it is the most accurate. Some researchers irrigate with wastewater, while others use regular water. This variation in irrigation water sources can aid in determining the differential responses of olive trees to varying water quality and availability scenarios.

• Watering strategies: The experiments involved various watering strategies, such as deficit irrigation that were either controlled or regulated. In contrast to deficit irrigation, control irrigation consists of supplying olive trees with the required water. This strategy permits researchers to investigate the effects of varying levels of water stress on the physiology, productivity, and adaptation mechanisms of olive trees.

• Geographic distribution: The majority of investigations have been conducted in Mediterranean countries, which is not surprising given the region's extensive cultivation that produced the world's 90% olives (FAOSTAT 2018). Nonetheless, it is important to note that a few experiments have been conducted in various parts of the globe.

As a whole, the diversity of experimental approaches reflects the complexity of investigating drought stress in olive trees and emphasizes the need for exhaustive studies under different conditions. This variation also highlights the significance of contemplating the experimental design, methodologies, and environmental factors when comparing and interpreting the results of different studies.

In response to research Q4, the analysis revealed the various impacts of climate change on olive cultivation, including erratic rainfall, high temperature, transpiration, spatial variability, and phenology. Transpiration in olives will decrease in response to climate change, with rainfed olives being affected the most. Further, the analysis shows that interannual variability in olive yield has increased due to water stress during flowering. Also, spatial variability is more pronounced in rainfed olive trees compared to irrigated ones. Moreover, it was found that flowering is a critical phase influencing olive yield, net profit margin, and irrigation requirements. Finally, there is a connection among phenology, irrigation management, and the damaging effects of heat and water stress events. These observations collectively demonstrate the complex relationship between climate, water availability, phenological stages, and the overall productivity and management of olive cultivation. The drought stress adaptation and mitigation strategies employed under climate change scenarios are as follows:

• High-efficiency irrigation: Promoting the use of efficient irrigation strategies for olives to manage water availability and reduce water loss.

• Selection of early flowering cultivars: Emphasizing the selection of olive cultivars that flower early, particularly for rainfed conditions, to mitigate the effects of climate change.

• Identification of new cultivation areas: Identifying and exploring new areas for olive cultivation that may be more suitable under changing climate conditions.

• Microclimate selection: Considering microclimatic factors when planning olive tree cultivation, such as choosing locations with favorable temperature and moisture conditions.

• Mulching: Implementing mulching techniques to conserve soil moisture and improve water retention in olive orchards.

• Soil management: Employing soil management practices to reduce runoff and improve soil water-holding capacity.

• Fertilization: Applying appropriate fertilization techniques, especially for young olive trees, to support their growth and adaptation to changing conditions.

• Cover crops: Using cover crops in olive orchards to enhance soil quality, moisture retention, and biodiversity.

• Olive ecosystems: This could perhaps help slow down climate change by storing CO₂ in tree biomass and soil, which is a natural way to deal with climate change. It thus helps with adaptation because the improved soil functions make it easier to store rainwater.

• The trees in the dry location exhibited an increase in root dry weight and a decrease in leaf number and relative stem height, whereas the trees in the wet location exhibited an increase in the leaf chlorophyll content and an increase in relative stem diameter and root hydraulic conductivity.

Drought stress affects the physiology, growth, yield, and development of olive trees. Single adaptation strategies are not enough to sustain Mediterranean olives, and both short- and long-term strategies are needed. Suitable areas for olive cultivation are shifting due to climate change, with changes in rainfall patterns and temperature variations affecting olive flowering. Olive trees exhibit different responses to drought stress based on their location, and they have developed mechanisms to cope with arid environments. Moreover, drought combined with stressors such as heat and UVB radiation does not worsen the negative effects of drought on olive trees. Furthermore, higher CO₂ concentrations increase the yield, while lower rainfall decreases it. Likewise, drought stress reduces CO₂ assimilation and chlorophyll content in olive trees. Overall, adaptation strategies, water management, and understanding olive tree responses to stress are crucial for mitigating the impacts of changing climate.

The effects of drought stress on olive trees have been the subject of a significant number of studies and research papers, which have been published as well as seen and studied earlier. In response to research Q5, the following are the key challenges and potential future directions deduced.

• Detection techniques: Given the advantages and limitations of various detection methods, future research could also focus on developing guidelines or decision support systems for selecting the most appropriate method based on research objectives, observational scope, and available resources.

• Diverse data sources: There is a need to consider the creation of frameworks that capitalize on the strengths of each data source while resolving their limitations.

• Remote sensing data: Data from remote sensing provides valuable insights into the health of vegetation and the availability of water across vast areas. Future research could concentrate on developing automated or semiautomated techniques for labeling remote sensing data, thereby reducing the amount of manual labor required.

• Standardization of experimental design: The considerable variability in the designs and methodologies of olive tree drought stress experiments poses a challenge in comparing and interpreting results. Future work could focus on developing standardized protocols and experimental designs to ensure consistency across studies.

• Cultivars assessment: The majority of the studies analyzed were limited to one or a few olive cultivars, so the results may not apply to other cultivars. Researchers should identify cultivars with higher drought tolerance and resilience, contributing to the development of more drought-resistant olive varieties.

In summary, future research should focus on comparative evaluations of detection techniques, including the exploration of deep learning with labeled datasets. Efforts should also be made to develop decision support systems for method selection and integrate diverse data sources.

Standardizing experimental designs and conducting long-term experiments are important for robust research, in addition to investigating the effects of different irrigation techniques and water sources, optimizing watering strategies, selecting resilient cultivars, exploring new cultivation areas, and conducting comprehensive research. Addressing these challenges and pursuing these research avenues would contribute to a more robust understanding of olive tree responses to drought stress and inform effective strategies for managing and mitigating its impacts. A detailed framework for olive tree drought stress monitoring and mitigation based on AI is proposed in Appendix A (available online).

This study presents an SLR of drought stress on olive trees under changing climates. The purpose of this article is to provide researchers with a detailed, comprehensive, and useful review of the current drought stress challenges and potential solutions. It emphasizes the methods, techniques, datasets, drought experiments, and potential future directions. The SLR analysis showed that researchers use in situ data for precise and detailed olive tree information, but these data have limited spatial coverage. Climate data have broad coverage but lack location-specific precision, while remote sensing data have large spatial coverage and quick monitoring but may lack tree-level detail. Combining these data types can help explain the olive tree drought stress better. Equally important, the diversity of experimental techniques evaluated shows the complexity of investigating drought stress in olive trees and the need for extensive investigations under different scenarios. The diversity also highlights the relevance of experimental design, techniques, and environmental aspects for comparing and evaluating future study results. Furthermore, this study proposes an innovative AI-based framework for monitoring and mitigating olive tree drought stress, providing valuable insights for sustainable management in the face of climate change challenges.

We are grateful to Euromed University of Fes, UEMF, Morocco, for providing the enabling environment to conduct this research seamlessly. We are also grateful to Google AI for supporting Kaloma Usman Majikumna with a Google Africa PhD fellowship. In addition, we appreciate Mr Oluwadamilare Harazeem Abdulganiyu and Dr Omar Eloutassi for their valuable feedback in improving this work.

K. U. M. and M. Z. conceived the concept for the article, K. U. M. conducted the literature search, drafting, data analysis, and synthesis, and M. Z. and A. H. A. analyzed the work critically and made necessary modifications where needed.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.