Introduction

Basal Stem Rot (BSR) disease is a major disease that is prevalent in oil palm plantations. Ganoderma is an economically important plant pathogen, especially in oil palm plantations, causing BSR disease. Ganoderma species are found in both tropical and temperate parts of the world’s green ecosystems (Rebitanim et al. 2020; Supramani et al. 2022; Pilotti 2005). Based on records, BSR disease has infected over 60% of the nation’s oil palm fields due to white-rot fungus of the Ganoderma spp. family (Naher et al. 2013). Specifically, it has been reported that at least seven types of species Ganoderma pathogen (G. boninense, G. miniatocinctum Steyaert, G. chalceum (Cooke) Steyaert, G. tornatum (Pers.) Bers., G. zonatum Murill, G. xylonoides Steyaert, G. ryvardense Tonjock and Mihand G. lobenense) associated with BSR disease can be found in Malaysia, Indonesia, Papua New Guinea, and Cameroon (Mih 2015; Rebitanim et al. 2020; Supramani et al. 2022). However, the most active and aggressive pathogen is the G. boninense species in Malaysia and Indonesia, causing BSR disease (Wong et al. 2012). The infected trees may die within six to 24 months for young palms and one to two years for mature palms once the symptoms are manifested since the fungus is not identified early and management measures are not implemented. In the context of Southeast Asian countries, specifically in Malaysia and Indonesia as the main oil palm producing countries, BSR disease outbreaks have been documented to predominantly occur in peatland regions, where they are influenced by the water table level, the pH of the peat at a depth of 0–15 cm, and the age of the oil palm trees (Supriyanto et al. 2020). Generally, the oil palm (Elaeis guineensis Jacq. Dura x Pisifera) hybrid is one of the main plantation crops in Malaysia and Indonesia, and these two countries together account for 85 to 90% of global exports (Swaray et al. 2020). According to Parveez et al. (2022), palm oil is also used as a feedstock for oleochemicals and a precursor for biodiesel. Malaysia reported a total export revenue of over $24.66 billion from palm oil and palm oil products. As the world’s top producers of palm oil, Malaysia and Indonesia need to immediately and efficiently address these issues (Wong et al. 2012; Bivi et al. 2016; Kamu et al. 2021). According to Paterson (2019), in his simulation for the evolution of BSR disease progress over the next 80 years using the Agricultural 4.0 approach, no significant innovations in the treatment of diseases were documented. Notably, some remediation techniques, such as developing oil palms in novel regions with suitable climates may help lower BSR disease progress. Moreover, between 2050 and 2070, the climate for oil palm will deteriorate dramatically in terms of cold, heat, dry stress, or drought occurrences, and BSR disease is expected to increase as a result of climate’s negative impact on oil palm development and resilience (Jazuli et al. 2022). As a result, there are around 25 years’ worth of possibilities to implement corrective measures to address future BSR disease levels.

As for resistance to BSR disease progression in seedling levels, the researchers evaluated the effectiveness of calcium carbonate on oil palm seedlings and discovered that pH 6 was the most efficient soil pH for suppressing BSR infection (Mahmud and Chong 2022). It is also recorded that a combination of silicon oxide, potassium silicate, calcium silicate, sodium silicate, and sodium meta-silicate decreased the severity of BSR in oil palm (Zakaria 2023). In fungicide treatment, hexaconazole was meticulously given to the damaged tree area by pressure injection and reported that 70% of treated oil palms produced fruit bunches after five years (Rashyeda et al. 2023). Moreover, application in pyraclostrobin successfully decreased BSR disease severity in oil palm seedlings while giving physiological advantages such as increased height, bole diameter, root mass, photosynthetic rate, PSII quantum efficiency, and relative leaf chlorophyll content (Said et al. 2019; Bharudin et al. 2022). In addition, injecting the fumigant fungicide dazomet into the stem lesion may limit the development of fruiting bodies. It was discovered that after 3 years of dazomet treatment ended, Ganoderma inoculum was discovered in infected stumps (Maluin et al. 2019). Despite the availability of certain efficient fungicides, their use in control methods may contaminate the environment and lead to the development of fungicide resistance (Shcherbakova et al. 2020).

In BSR disease assessment in oil palm plantations, the assessment methods of BSR disease are still ongoing research as various techniques have been applied to manage early or moderate stages of infection, including destructive or non-destructive methods. The most basic assessment techniques and the most accurate have been applied through the manual inspection of individual trees (naked-eye assessment) by the farmers (Kurihara et al. 2022). In a different method, the farmers need to cut down the tree (destructive method) to inspect the disease in detail (lab work). Other than that, both methods are not preferable as they require high labour, costs, and time, especially for large-scale oil palm plantations. In non-destructive assessment, other researchers looked into an alternative using geospatial technologies and other advanced instrumentation (Multispectral data, Hyperspectral data, Terrestrial Laser Scanning (TLS), Intelligent Electronic Nose (e-nose), Tomography, and Radio Detection and Ranging (RADAR) (Hamidon and Mukhlisin 2014; Kresnawaty et al. 2020; Hashim et al. 2021; Khaled et al. 2021; Baharim et al. 2023. Overall, most of them received a positive and significant result in their research work. Based on this motivation, this research study tends to analyse in different ways through leaf physiologist variable specific water use efficiency (WUE) as it can provide a significant indicator for plant health status (stress condition) and helps for better understanding of BSR disease.

Assessment and application of WUE for oil palm studies have been widely explored in various aspects, including physiology changes in oil palm due to many factors, such as stress conditions and genetics. In oil palm management, appropriate irrigation is capable of improving oil palm plantations significantly. Eshkaftaki and Rafiee (2020) developed models to estimate crop yield by utilising the WUE variable, revealing a satisfying model. In different research, the WUE variable has been used as the main parameter in analysing the effective application of bio-silicic acid (BioSilAc) during periods of low rainfall (limited water supply). The result indicated the bio-silicic acid in mature oil palm grown at 75−100% NPK + 4 tablets were able to improve yield at 11.9% (Treatment 5) and 12.1% (Treatment 4) as well as water use efficiency at 31.3% (mature) and 50.4% (immature) compared to the control treatment (Santi et al. 2021). In different irrigation system assessments, WUE data were fully utilised to determine the most significant method of irrigation to be applied in oil palm plantations (Rao 2009). According to Roslan et al. (2013), WUE can be enhanced by reducing the evapotranspiration rate and vapour pressure deficit by producing high WUE and drought-tolerant oil palms through breeding, reducing surface runoff, and maintaining soil water availability. Combining drip irrigation with fertilizer application can also effectively improve WUE in oil palms (Manorama et al. 2020). In nursery management, WUE parameters have been adopted to analyse oil palm seedling growth with varying variables exposed to short-term CO2 enrichment in a closed top chamber with good results. The management of oil palm nurseries can be improved by enhancing the early development and biomass of oil palm seedlings by CO2 enrichment (Ibrahim et al. 2010). In analysis for oil palm seedlings growth under water stress condition, it was revealed by Syaripah Najihah et al. (2019) that severe water stress affect the WUE variables and other leaf physiologies (vegetative plant growth, photosynthesis rate, transpiration rate, stomatal conductance, leaf water potential, relative water content, and leaf moisture content) significantly.

Despite the enormous potential involving WUE parameters in oil palm research for plantation management and plant physiological changes due to stress or different factors, there is limited information and a remarkable research gap in oil palm cultivation involving WUE in oil palm tree reactions to plant diseases. Based on this literature review, no previous studies have reported on the assessment of WUE performance or application on WUE parameters and plant disease, specifically for BSR that threatens oil palm sustainability. Therefore, this study explored the correlation of WUE variables with BSR disease severity for mature oil palm trees by focusing on the WUE leaf-level scale in different frond levels (9- young and 17- mature) as a less destructive measurement method to elucidate the relationships between these variables and proposed a new model from WUE variable for effective assessment in BSR disease severity in oil palm plantation.

Materials and methods

Selection study area

The study was conducted in the Perak state of Malaysia. Specifically, the area chosen in Perak was Lekir, which was located on the West Coast of Peninsular Malaysia with coordinates of 4°11′51.89 N, 100°47′11.92″ E for latitude and longitude, respectively. Overall, for the chosen area, the region was dominated by agricultural businesses, mainly oil palm plantation that has been developed by Felcra Lekir Berhad to assist the Government of Malaysia as one of the country’s commodities. Given the local farmers’ concerns regarding BSR disease and the issues raised by the Malaysian Palm Oil Board (MPOB) research agency about the incidence of the disease in their oil palm plantations, further assessment of the disease was required. Regarding the breed and soil type of oil palm in plantations, this research solely focused on the Dura and Pisifera (DxP) hybrid type in coastal soil since it was widely planted in Malaysia’s commercial oil palm sector as in Lekir, Perak, Malaysia.

In terms of adequate water and nutrient supplies, the local farmers in the chosen region provided sufficient and timely fertilizer together with the establishment of water flow (irrigation) for their oil palm plants. The area chosen was also disease-free from other known diseases or pest infestations.

Manual inventories of oil palm trees for Basal Stem Rot (BSR) disease severity

The ground census data was collected on the site to support the analysis and assisted by the BSR disease team of experts from the MPOB as revealed in Fig. 2(c), following a standard BSR severity levels guidelines as shown in Table 1. Figure 1 depicts the sample of trees in the oil palm plantation according to the BSR disease severity levels for healthy (T0) and non-healthy tree (T1, T2, T3). In the healthy category (T0), the frond view characteristic looked healthy and green, with most of the frond number existing at the oil palm crown tree with no broken or yellowing frond. Besides that, all frond crowns opened widely on the tree. Next, in the early infection categories (T1), the characteristic of the crown tree showed little change compared to the healthy crown tree (T0) except only 1 or 2 fronds crown tree starting to unopen and turn yellow. In detailed observation of the tree trunk, small white button of Ganoderma started to appear, and some were already mature as shown in Fig. 2(a). However, the characteristic of the crown tree started to be obvious in the moderate infection (T2) whereby more frond crown trees were changing colour (yellowing) and normal structure (bending down) with less fruit produced together with mature Ganoderma appearing either in single or many around the trunk as depicted in Fig. 5(b) and some of it rotten at bottom trunk and create a big hole on it as shown in Fig. 3(a) and (b). Finally, in the final view of BSR disease (T3) characteristic, the frond crown tree lost half of its crown on the ground (falling) with only a few lefts on the tree. Moreover, the remaining fronds on the tree also turned yellow and brown and were ready to fall; and no more fruit was produced from the tree. Moreover, some of the sample trees in T3 were smaller compared to the others according to the ground census.

Fig. 1
figure 1

Overview oil palm tree condition different BSR disease severity (T0 - Healthy, T1- Low, T2 - Moderate, T3- High)

Full size image
Fig. 2
figure 2

(a) Small white button of Ganoderma appears (b) Many mature Ganoderma appears (c) Manual inventories for Ganoderma on an oil palm tree

Full size image
Fig. 3
figure 3

(a) Rotten at bottom trunk more than 50% (b) Big hole appears to bottom trunk of the tree that fit whole foot inside of it

Full size image
Table 1 Shows the guidelines by MPOB for BSR disease severity in mature oil palm
Full size table

Sampling methods techniques in leaf gas exchanges for WUE and chlorophyll contents

In this sampling method, the sample trees were selected from four levels of BSR disease severity (T0, T1, T2, T3) in 6 trees for T0, 6 trees for T1, 6 trees for T2, and 6 trees for T3. This amounted to a total of 24 trees for oil palm tree samples. For healthy oil palm growth, the number of fronds can usually reach up to 40–50 frond number and even more (Aholoukpè et al. 2013). Two frond numbers were choosing (frond 9 and frond 17) for WUE (E) (µmol CO2 H2O mmol−1) measurement with other variables in response to BSR disease severity due to the limitation of time. A total of 48 cut fronds samples were collected for 24 trees of oil palm tree as shown in Fig. 4(a). The main criteria for choosing frond 17 as a sample frond number was due to its position in the centre of the frond in oil palm tree and it has been reported as the best frond number for nutritional evaluation based on early findings (Jayaselan et al. 2017; Ayanda et al. 2020; Yadegari et al. 2020). Meanwhile, frond 9 was chosen as a sample along with frond 17 since this frond was regarded as a young frond level in oil palm trees.

Fig. 4
figure 4

(a) Selection samples from frond number 9 and 17, (b) Cut leaflet from rachis parts of frond sample (3 left and 3 right), (c) WUE measurement using LiCOR instrument, (d) Chlorophyll content measurement using SPAD instrument for leaflets

Full size image

Therefore, this research provided a new analysis perspective on two different frond levels (young and mature frond levels) for comparative assessment between WUE performance and BSR disease severity. In the leaflets sample from both frond numbers, only six (6) leaflets were selected from the whole leaflets. The leaflet samples were obtained from fronds started from the rachis part as recommended by the MPOB guidelines (Fig. 4(b) in which accumulated 12 reading for 1 sample tree. This method was applied to all frond samples in this research. Thus, the number of samples from leaflets in this study for WUE and BSR disease severity was 288 leaflets for 48 frond samples of oil palm trees. Leaflet samples measurement of leaf gas exchange parameters was also recorded consisting of Photosynthetic Rate (Pr), Transpiration Rate (Tr), Temperature Leaf (TempL), Intercellular CO2 Concentration (Ci), and Stomata Conductance (gs) using two portable equipment sensors of photosynthesis system (LI-6400XT, LiCOR Inc., USA) while the leaflet was still intact with the frond sample for both fronds number as shown in Fig. 4(c). Before a measurement was taken, the leaflets samples were checked to be in good condition to ensure good quality data: clean without any dust or any particles on leaflets (the leaflets being cleaned with clean tissue), no broken leaflet, or no obvious unhealthy spot on leaflets. The condition of the inside of the leaf curvette for LiCOR was as follows: ambient humidity, ambient temperature, photosynthetically active (PAR) with 1000 photon µmol m−2 s−1 (red/blue light) and reference CO2 of 400 µmol mol−1. Meanwhile, the main parameter in this study involved the calculation of the instantaneous WUE (E) variables using the formula WUE = (Pr/Tr) (A’fifah et al. 2018). The data measurement was conducted during the photosynthesis period between 8.00 and 11.30 a.m. (Koyama and Takemoto 2014). All the data variables collected were automatically stored in LiCOR and downloaded for further analysis.

The data collection continued with chlorophyll content (F) reading on site right after measurement LiCOR finished. The leaflet has been cut out from the frond sample and the reading leaflet was carried out using a portable Soil Plant Analysis Development (SPAD 502) as shown in Fig. 4(d). The leaflet was measured at three (3) separate locations (left, centre, and right) on both sides, yielding a total of six (6) measurements that were automatically averaged. Data were manually stored, transferred, and then used in the subsequent study. At the end, all the chlorophyll data also have 12 readings for each sample tree (24 samples) with a total of 288 sample leaflets.

WUE model development in estimation of BSR disease severity

The research work continued with the development of an empirical model with WUE variables in predicting BSR disease severity in oil palm plantations from averaged samples in fronds 9 and 17. The model development utilized a Stepwise Multiple Linear Regression model from software (IBM SPSS 20 version, USA) software in which specializes in structural equations modelling development (Liu et al. 2021). In order to produce a stable and reliable model, the cross-validation dataset is applied in splitting into 70% for training and 30% validation model generated (Kee et al. 2023).

Statistical analysis in WUE and BSR disease research

In this study, the statistical analysis was conducted utilizing IBM SPSS software which is similar to model development that also specialises in understanding data, analysing trends, forecasting, and planning to validate assumptions and drive accurate conclusions. For WUE performance with BSR disease severity analysis, this research used Whisker box-plot graph analysis to identify the patterns between these variables. In a second analysis for correlation, the research work focused on the Pearson Correlation between WUE and BSR disease severity in different frond levels. In addition, the Multiple Linear Regression (MLR) model was applied to determine the significant variables contributing to WUE performance. Finally, in the model performance developed, the research applied Root Mean Square Error (RMSE) in identifying error differences between actual WUE on the ground and with the model forecast generated (Hodson 2022). The formula for RMSE revealed as above:

$${\text{RMSE}} = \left[ {\frac{1}{n}\mathop \sum \limits_{i = 0}^n \left( {{P_{i}} - {O_i}} \right)} \right]\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/\kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} $$

Where \(\:{P}_{i\:}\) is set of model data predictions (i = 1, 2,…, \(\:n\)).

  \(\:{O}_{i}\) is the actual observation for each \(\:{P}_{i\:}\)

Results and discussion

Table 2 revealed the WUE data measured on the ground while Fig. 5 presents the result of WUE versus BSR disease severity levels using the Whisker box-plot graph to elucidate the pattern. This Whisker box-plot graph revealed that each of the disease severity of fronds 9 and 17 starts from T0, T1, T2, and T3.

Fig. 5
figure 5

Whisker box-plot graph for WUE and BSR disease severity in fronds 9 and 17

Full size image
Table 2 WUE values in frond 9 and 17 collected (24 samples)
Full size table

Analysis of WUE performance with BSR disease severity in frond 9 and frond 17 for mature oil palm tree level

In this study, the performance of WUE and its relationship with BSR disease severity in different frond levels (frond 9 and frond 17) were analysed as shown in the Whisker box-plot in Fig. 5. Based on Fig. 5, the result presented a reliable pattern between these variables. The value for healthy level (T0) oil palm categories was almost similar between frond 9 and 17, especially from frond 17, which had an overall high-level value of WUE, especially in the first quartile (25%) of the Whisker box-plot that ranged from 2.5 µmol CO2 H2O mmol−1 to 4 µmol CO2 H2O mmol−1. Since no previous study has attempted a similar investigation for mature trees, this study identified several other reliable WUE data with healthy conditions for oil palm trees of different ages of oil palm, breeds and types of plants that shared similar photosynthesis characteristics as oil palm (C3 category) using WUE (leaf level) assessment.

The findings of WUE for the healthy oil palm was similar to a study by A’fifah et al. (2018) in which a minimum value of 2.6 µmol CO2 H2O mmol−1 was recorded in May 2012 (the same month when the present study was conducted) of DxP breed (the same breed as in this study). In different clones of oil palm research by Jazayeri et al. (2015) for healthy oil palm, the WUE values also ranged from 2.4 µmol CO2 H2O mmol−1 to 3.1 µmol CO2 H2O mmol−1. Furthermore, the findings were similar to Suresh et al. (2012) who investigated tenera hybrid of the healthy oil palm trees and reported WUE values ranging from 3.6 µmol CO2 H2O mmol−1 up to 4.3 µmol CO2 H2O mmol−1. In different WUE assessments for healthy plants of grapevines (C3 plant photosynthesis category), the WUE assessed by Medrano et al. (2015) ranged from 2 µmol CO2 H2O mmol−1 up to 5 µmol CO2 H2O mmol−1. Based on a meta-analysis conducted by Flexas et al. (2016) for the WUE level of 119 species from C3 plant categories, the value ranged from 3.5 µmol CO2 H2O mmol−1 to 6 µmol CO2 H2O mmol−1 for healthy plants. In conclusion, it can be justified that the result obtained between WUE and BSR disease is reliable for healthy plants (oil palm trees) not affected by any other variables, such as water or heat stress. In terms of outliers existing in both of frond 9 and 17 (Fig. 5), for the T0 healthy category, this result may be linked to the fact that some of the leaflet samples were in a very good condition. According to Lambers et al. (2008), the value can reach up to 11 µmol CO2 H2O mmol−1. For the T1 category in both frond 9 and 17 for the first quartile (25%), the WUE value was almost similar between both fronds; however, it slightly decreased when being compared with T0 level of WUE in the first quartiles. Moreover, the second quartile (25%) saw 4 µmol CO2 H2O mmol−1 for frond 9 and 3.5 µmol CO2 H2O mmol−1 for frond 17, both of which were median levels with significant differences for WUE, especially from frond 9 compared to frond 17. This was probably due to frond 9 and 17 still having good conditions of photosynthesis and transpiration processes despite being in the early infection of BSR disease. The leaflets from both fronds indicated little changes from field observation as presented in the WUE value, especially in the frond 9. For the T2 category, the WUE level from Whisky Box-Plot for both fronds revealed at the second quartile (25%), a decrease for frond 9 and 17 from 4 µmol CO2 H2O mmol−1 H2O to 1.8 µmol CO2 H2O mmol−1 and 3.5 µmol CO2 H2O mmol−1 to 2.5 µmol CO2 H2O mmol−1, respectively. This result was expected as the BSR disease infection severity increased, and the oil palm tree might be affected by water stress since the BSR disease affected the root system of the tree. For the T3 category, both fronds showed that the second quartile (25%) was more obvious at frond 9 and frond 17. The median of both levels for WUE was less than 2 µmol CO2 H2O mmol−1 with minimum values of 0.5 µmol CO2 H2O mmol−1 and 0 µmol CO2 H2O mmol−1. The BSR disease severity showed a significant change in visual appearance as compared to the other severity levels (T0, T1, T2) as the frond itself started to bend down more than 50% (T3) with colour changing to yellow or even brown in some leaflets. Moreover, some of the T3 samples from the frond 17 had 100% bending crown close to the tree trunk.

Table 3 Pearson Correlation Matrix
Full size table

Correlation analysis between WUE with different variables for Frond 9 and Frond 17 according to BSR disease severity

Table 3 of the Pearson Correlation Matrix summarizes the correlation between WUE variables of Photosynthetic Rate, Transpiration Rate, Temperature Leaf, Intercellular CO2 Concentration, Stomata Conductance, and Chlorophyll Content for both fronds. This analysis focused on the correlation between variables that were more than −0.5 and 0.5. (9 and 17). For frond 9, the variables of photosynthetic and transpiration rates had positive correlation with the WUE variables, with the values between 0.506 and −0.617. However, a lower correlation values were identified for frond 17 with the values of 0.487 and −0.113, respectively (Table 3). This result showed that photosynthesis was positively affected by the BSR disease, especially for T2 and T3 categories, and was significantly correlated with WUE. In contrast, negative correlations were identified between WUE and transpiration rate as the transpiration rate reacted inconsistently for the different severity levels as compared to the photosynthesis. This condition was also similar between the WUE variables, with intercellular CO2 and stomata conductance variables showing values of −0.892 and −0.536 and a decrease in frond 17 by −0.833 and −0.135, respectively. Most of the WUE variables were correlated with different variables, showing that the changes in plant physiology for both frond 9 and 17 were related to BSR disease severity.

In term of the photosynthetic rate correlation with transpiration rate, the result indicated the value of 0.208 for frond 9, but an increase in frond 17 to 0.703. It showed that the BSR disease significantly reduced photosynthesis rate, especially in the T2 and T3 categories and some of the leaflets from the fronds were broken and lost. Meanwhile, the transpiration rate reacted similarly (i.e., decreasing values) as photosynthesis since the BSR attacked the plant root system and blocked the water uptake to the tree (Baharim et al. 2021), causing stress condition to the tree. Transpiration rate efficiency declined to reduce the limit of water loss during the transpiration process (Li et al. 2023). In the aspect of the photosynthetic rate variable with leaf temperature, the value was less correlated in frond 9 with −0.259, but more correlated in frond 17 with −0.513. For the normal plant growth, photosynthesis reacts positively to the temperature value at an optimum degree (Huang et al. 2019). Based on the raw data collected in the optimum leaf temperature (28°- 31 °C) for all BSR severity levels, it was correlated with the photosynthesis rate in negative value. In general, for stomata conductance variable with photosynthetic rate, the stomata was expected to be open during the photosynthesis process (Kusumi et al. 2012). However, in this scenario the contrary occurred since it was affected with BSR disease, and the photosynthetic rate variable was correlated with stomata conductance in frond 9 despite the value of 0.237 being less than frond 17 (0.663). Since the photosynthetic rate decreased with the increasing BSR severity, the stomata conductance was also positively correlated. The changes in BSR severity, especially from T2 and T3 led to this result as most of the fronds started to bend down close to the trunk and some of the leaflets from this frond already changed their colour to yellow or brown and wilting. The changes in these features (morphological and biochemical changes) influenced greatly and led to reductions of these variables (Luong and Loik 2022). This result aligned with Verma et al. (2020) who concluded that lower photosynthetic rate might occur by deteriorating photosynthetic pigments, resulting in total failure for the crop. In the context of transpiration rate with temperature leaf, frond 9 showed a low correlation of 0.113, but a high negative correlation was detected in the frond 17 (−0.569), showing that frond 17 was more correlated since the temperature leaf affected the transpiration rate but in negative value as compared to frond 9.

In transpiration rate correlated with intercellular CO2 concentration, a high correlation in frond 9 with 0.646 as compared to the frond 17 with 0.222 was observed. This result indicated that while the transpiration rate decreased, the intercellular CO2 concentration also decreased. Furthermore, the transpiration rate correlation assessment of both fronds (9 and 17) with the stomata conductance revealed high correlation at 0.947 and 0.982, respectively. This result was expected since as the transpiration process might experience water stress due to BSR disease and eventually limit their process, the same also occurred for stomata conductance reactions to this condition (Yin et al. 2020). The correlation between temperature leaf and stomata conductance revealed that frond 9 had very low correlation (−0.133) while frond 17 had high correlation (−0.575). This research concluded that frond 17 received more sunlight (wider frond) as compared to frond 9 (in between the other oil palm frond number), especially from healthy (T0), and low (T1) levels BSR disease categories in terms of canopy structure while the correlation between intercellular CO2 concentration with stomata conductance showed high correlation in frond 9 (0.599), but low value in frond 17 (0.240). This result was interpreted that as stomata conductance was regulated slowly to close by BSR severity infection due to increasing stress, the intercellular CO2 concentration also declined in both frond levels (Flexas and Medrano 2002).

Tables 4, 5 and 6 reflect the output obtained from MLR models to determine the significant variables contributing to WUE data based on different frond levels (fronds 9 and 17). According to Table 4, the model summary depicted that the Adjusted R Square values for both frond levels (9 and 17) explained 0.863 (86.3%) and 0.826 (82.6%) of the variation in WUE data, respectively. Table 5 depicts the model fit (F-test) in which both fronds (frond 9 and 17) satisfied the model at 0.000 and 0.000 significant levels and p-value < 0.05. Both fronds also demonstrated that the most contributing or significant variables (either negative or positive) for WUE were Photosynthetic Rate (Pr), Transpiration Rate (Tr), and Intercellular CO2 Concentration (Ci) variable which corresponded to beta values of 0.495, −0.466 and −0.349 for frond 9 while 0.605, −0.369 and −0.533 for frond 17, respectively.

Table 4 Multiple Linear Regression model summary for frond 9 and frond 17
Full size table
Table 5 Analysis of variance (ANOVA) for frond 9 and frond 17
Full size table
Table 6 Coefficient table for frond 9 and frond 17
Full size table

Determination of significant variables contributes to WUE performance in mature oil palm level

In the third analysis to identify the significant variable that contributed to WUE (E) in different fronds (9 and 17), six (6) variables were analysed using the Multiple Linear Regression (MLR) models, including the Photosynthetic Rate (Pr), Transpiration Rate (Tr), Temperature Leaf (TempL), Intercellular CO2 Concentration (Ci), Stomata Conductance (gs), and Chlorophyll Content (F). The results (Table 4) of the model summary from frond 9 and 17 indicated that the variance for both models in high-level value (more than 0.5) for Adjusted R Square (0.863 and 0.826, respectively). At this point, frond 9 had higher than frond 17 for model variance based on various plant physiology variables assessment. This result also supported the ANOVA test that indicated both models from fronds 9 and 17 were in significant level (0.000) since the p-value < 0.05 for acceptable model fit (Table 5). To determine which variable contributed the most to the WUE data, Table 6 revealed that frond 9 and frond 17 had similar values for Photosynthetic Rate (Pr) (−0.495, −0.605), Transpiration Rate (Tr) (−0.466, −0.369) and Intercellular CO2 Concentration (Ci) (−0.349, −0.533), respectively. At this point, it was concluded that Photosynthetic Rate (Pr) and Transpiration Rate (Tr) were the main contributors based on the formula measurement of WUE in leaf concentration (Pr/Tr), and Intercellular CO2 Concentration (Ci) variable also indicated good performance and could be considered in further analysis for WUE data. Other than that, this variable could also be used as a model variable for model prediction.

Table 7 Descriptive statistics for WUE modelling estimation BSR disease severity
Full size table
Table 8 Generated summary model in WUE modelling estimation BSR disease severity
Full size table
Table 9 Coefficient model table for frond 9 and 17 in WUE modelling estimation BSR disease severity
Full size table
Table 10 Summary empirical model derived for WUE modelling estimation BSR disease severity
Full size table
Table 11 Result accuracy assessment in training model sample WUE modelling with Root Mean Square Error (RMSE)
Full size table
Table 12 Result accuracy assessment in validation model sample WUE modelling with Root Mean Square Error (RMSE)
Full size table

Assessment for WUE modelling in BSR disease severity

In final research work for this study, a new model has been developed in assessment for BSR disease severity based on the dataset (averaged frond 9 and 17) in previous analysis. The model was developed using Stepwise Multiple Linear Regression in SPSS software for all variables used with cross-validation splitting dataset (70% = training model, 30% = validation model) and generated a few tables for model development includes descriptive statistic for training and validation model dataset (Table 7), summary in model (Table 8), and coefficient table model (Table 9). However, to make it easier and simple in view for model developed, Table 10 summarized the models equation accordingly. Throughout the models generated, the result demonstrated in Tables 11 and 12 that all models provide a significant result for training and validation model in assessment BSR disease severity with ranged RMSE (0.54–0.19, 0.48–0.32) respectively. In terms of rank order, the model is sorted in Model 1, Model 2, Model 3, and Model 4 in ascending order based on RMSE accuracy assessment. However, in Model 3 and Model 4 for training and validation RMSE, it’s revealed that do not have much difference between them. In individual comparison difference for BSR disease severity (T0, T1, T2, T3) on actual and the model forecast, the result demonstrated that all of them in range (0.00–1.84). Thus, these new models contributed to strengthening the body of knowledge in application of WUE variable in assist for assessment BSR disease severity in plantation. Other than that, this result also highlighted the use of Stepwise Multiple Linear Regression method as model development selection as it capable to handles large dataset in various variables and generated a good prediction model in plant disease specific in BSR disease.

Conclusion

Prolonging the oil palm age to obtain the longest period with significant production is the most important concern among farmers. The productivity of oil palm declined in large amount through disease infection, especially from Basal Stem Rot (BSR) disease even in the immature trees. Analyzing water use efficiency (WUE) as the main variable in trees with BSR disease infection can effectively provide insights for fundamental knowledge in plant physiology reaction together with plant disease. Ultimately, this research is expected to significantly contribute and impact the plant disease (BSR disease) research and interactions with oil palm plant growth (plant physiology).