Introduction

It is ecognised that urban green spaces provide a range of social, health, economic and environmental benefits that improve the liveability of cities and thus human well-being (Tzoulas et al. 2007; Kondo et al. 2018; Marselle et al. 2020). For example, the role of urban green spaces in mitigating the urban heat island effect has been well documented (Sun and Chen 2017; Balany et al. 2020). However, there are many challenges associated with the development and expansion of urban green spaces (Haaland 2015; Aronson et al. 2017). Of these challenges, the mitigation of climate change-associated extreme climatic events will be amongst the most difficult (Zhang et al. 2019; Hobbie and Grimm 2020; Ramyar et al. 2021).

It is projected that the frequency and intensity of drought events will increase under climate change (Seneviratne et al. 2012; Kim and Jehanzaib 2020). The stress imposed on plants by these events can facilitate a range of morphological, physiological and biochemical changes, and even mortality in severe cases (Seleiman et al. 2021). The detrimental impacts of drought have been widely observed in both natural (Young et al. 2017; Hartmann et al. 2018; Hammond et al. 2022) and urban forests (Dale and Frank 2014; Kagotani et al. 2016; Nitschke et al. 2017; Bialecki et al. 2018). Urban forests are likely to be more vulnerable to the impacts of drought than forests with exposed soil surfaces (e.g. natural forests, plantations) due to the impervious surfaces of urban areas reducing rainfall infiltration (Mejía and Moglen 2010; Yao et al. 2016; Shao et al. 2019). In fact, it has been suggested that drought is the most significant factor limiting urban greening globally (Clark and Kjelgren 1990). This, coupled with the fact that at least 79 cities globally have experienced extreme drought events in the last two decades (Zhang et al. 2019), means it is vitally important that the drought resilience of urban forests is enhanced to ensure they continue to perform key ecosystem services (e.g. air filtering, micro-climate regulation, noise reduction; see Bolund and Hunhammer 1999).

Several strategies can be incorporated into the management of urban forests to enhance their resilience to drought, including supplementary irrigation (Livesley et al. 2021), application of soil amendments and biostimulants (Manea et al. 2023), mulching (Chalker-Scott 2007) and selection of drought-tolerant plant species (Sjöman and Nielsen 2010; Sjöman et al. 2015). However, many of these strategies are expensive and labour-intensive to implement. Of these strategies, biostimulant addition is amongst the most cost-effective and viable in terms of labour demand to enhance urban forest resilience to drought.

Biostimulants contain a diverse range of extracts obtained from organic raw materials that contain bioactive compounds (Du Jardin 2015). Biostimulants can be applied to plants by seed soak, root drenching, foliar spray or soil application and have been shown to improve plant growth, nutrition and stress tolerance (Calvo et al. 2014; Halpern et al. 2015; De Pascale et al. 2018; Li et al. 2022). Some of the most common biostimulants are derived from humic acid, protein hydrolysate and seaweed (Du Jardin 2015). Humic acid biostimulants enhance cell turgor, water absorption and reactive oxygen species detoxification (Canellas et al. 2015), protein hydrolysate biostimulants stimulate nitrogen assimilation, which improves carbon assimilation and photosynthate translocation (Colla et al. 2017) and seaweed biostimulants enhance water use efficiency and stimulate root growth (Ali et al. 2021). However, in many cases, the underlying mechanisms by which biostimulants perform these functions to reduce drought stress are still unclear (Ma et al. 2022). To date, research on the use of biostimulants in mitigating drought stress in plants has primarily focused on agricultural and edible horticultural plants and has reported a positive effect (Van Oosten et al. 2017; Bulgari et al. 2019). In contrast, the effectiveness of biostimulants in mitigating drought stress in urban plant species is less predictable, with both positive effects (Ferrini and Nicese 2002; Fraser and Percival 2003) and no effects (Fraser and Percival 2003; Banks and Percival 2014) being reported, suggesting further research is required on these species.

The aims of this study were to determine whether biostimulant application (1) improves the performance and (2) mitigates drought stress in plant species frequently planted in Australian urban areas. We hypothesised that:

  1. 1.

    Biostimulant application would enhance plant performance irrespective of watering treatment.

  2. 2.

    Drought stress would reduce plant performance but biostimulant application would mitigate this impact.

Materials and methods

Species selection and plant preparation

We selected six evergreen woody plant species and three graminoid plant species that are commonly planted in Australian urban areas for the study. The species were selected based on their popularity amongst growers in Australia (Plummer unpublished; Table 1) and stock availability at the time of the study. All the species are considered to be fast-growing and moderately drought-resilient, according to nursery industry classifications (Tabassum et al. 2021). For each species, 61 similar-sized plants grown in 0.25 L containers were purchased from a commercial nursery (Burringbar Rainforest Nursery, Upper Burringbar, NSW, Australia). The woody and graminoid plant species were grown for ~ 20 and ~ 12 wks respectively by the commercial nursery before purchase.

Table 1 Taxonomy, growth form and number of growers in Australia for each study species
Full size table

Within each species, the plants were arranged in order of size based on plant height for the woody plant species and leaf number for the graminoid plant species. Once arranged, five plants across the size spectrum were selected for the pre-treatment harvest, while the remaining 56 plants were evenly split based on size into the eight treatments (i.e. seven plants per treatment; treatments described below). The plants were then bare-rooted and transplanted into 3.2 L cylindrical containers containing 3 L of organic potting mix (Australian Growing Solutions, Tyabb, VIC, Australia). Approximately 5 g of 16 N:4.4P:8.3 K slow-release fertiliser (Nutricote, Yates, Padstow, NSW, Australia) was homogenised into the potting mix of each container. The plants were then allowed to establish in climate-controlled glasshouses (environmental conditions described below) for 10 wks. During this establishment period, the plants were mist watered thrice daily for 5 min.

Pre-treatment harvest and experimental design

After the establishment period, the growth metrics of each plant were measured. For the woody plant species, plant height and stem diameter were measured while for the graminoid plant species, leaf number was counted. Plant height was measured from the soil surface to the apical meristem, while stem diameter was measured at 15 cm above the soil surface using a caliper. After the measurements were completed, the pre-treatment harvest plants were harvested, oven-dried at 70 °C for 48 h and weighed using an analytical balance (Mettler Toledo, Port Melbourne, VIC, Australia). The pre-treatment harvest data were used to estimate the initial biomass of the treatment plants. For each study species, the biomass data of the pre-treatment harvest plants were plotted against the plant height (woody plant species) or leaf number (graminoid plant species) data to generate a straight-line equation. Using this equation, the pre-treatment biomass of the treatment plants was estimated based on their pre-treatment plant height or leaf number.

The treatment plants were grown in a fully factorial glasshouse experiment with two treatments: watering (drought-stressed, well-watered) and biostimulant (control, humic substance, protein hydrolysate, seaweed). This gave a total of 504 plants (9 species × 2 watering treatments × 4 biostimulant treatments × 7 replicates). The plants allocated to the well-watered treatment were connected to a drip irrigation system, which delivered 100 mL of water through a spike, three times a day. The drought stress treatment was imposed using a method developed by Marchin et al. (2020) that lowers volumetric soil water content (VSWC) to the desired level and then maintains it over time. This method involved placing the plants on top of floral foam blocks placed in tubs containing water, which the plants draw up using capillary action. The water was gradually drawn down over a 3-wk period until no water remained in the tubs. VSWC was subsequently maintained at ~ 3% for 12 wks (i.e. the treatment period). Leaf wilting and dieback typical of drought stress were regularly observed in the study species at this VSWC. The VSWC of each container was measured thrice weekly using a Hydrosense II Portable Soil Moisture System (Campbell Scientific Australia Pty Ltd, Garbutt, QLD, Australia). Whenever VSWC dropped below 3%, 50 mL of water was added to each container to prevent the plants from dying.

The biostimulant treatments were applied fortnightly during the 12-wk treatment period (i.e. five applications). For each application, the plants received 50 mL of biostimulant solution by soil watering. The biostimulants used were: (1) humic substance (5 mL/L; Liquid Humus, Nutri-Tech Solution, Yandina, QLD, Australia; Active ingredient: Potassium humate); (2) protein hydrolysate (5 mL/L; Hydrofish, No Frills Fertilisers, Margaret River, WA, Australia; Active ingredient: Fish waste; 3.1 N:0.34P:0.34 K); and (3) seaweed (SW; 6.25 mL/L; Seasol, Seasol International, Boronia, VIC, Australia; Active ingredient: Kelp; 0.1 N:0.01P:1.5 K). These solution concentrations were per the manufacturer’s recommendation. The plants in the control treatment received 50 mL of water.

The experiment was conducted in the Macquarie University Plant Growth Facility. For each species, each watering × biostimulant treatment combination was evenly split between four climate-controlled glasshouses. The plants were rotated through the glasshouses fortnightly to ensure there was no glasshouse effect. For the duration of the experiment, the glasshouse temperature was set to 25/19°C day/night and was continuously maintained by a fan coil unit using a water cooling and heating system. The humidity (75 ± 10% at 1400 h) and photosynthetic active radiation (827 ± 100 µmol m− 2s− 1) of the glasshouses were monitored continuously using a multi-grow controller system (Autogrow Systems, Auckland, New Zealand). When photosynthetic active radiation dropped below 800 µmol m− 2s− 1 between 0800 and 1700 h, supplementary light was provided to the plants to maintain photosynthetic active radiation at 800 µmol m− 2s− 1. As needed, all the plants were sprayed with white oil (Brunnings, Oakleigh South, VIC, Australia) for pest control.

Assimilation rate measurements

The assimilation rate of the plants was measured after the last biostimulant application. The assimilation rate was not measured for O. japonicus due to its small leaf size. For each plant, the assimilation rate was measured on three randomly selected new fully expanded outer-canopy leaves using a LICOR-6800 portable photosynthesis machine (LICOR Biosciences, Lincoln, NE, USA). The temperature (25 °C), relative humidity (75%), photosynthetically active radiation (500 mol m− 2s− 1) and CO2 concentration (450 ppm) within the machine’s cuvette were kept constant between measurements. The measurements were conducted between 0800 and 1300 h when the plants were the most photosynthetically active. The measurements were conducted in seven blocks, with each block containing a plant from each species for each watering × biostimulant treatment combination. Within each block, plants were randomly selected for measurement to ensure no measurement time effect existed among the watering × biostimulant treatment combinations.

Growth metric and allocation measurements

After the 12-wk treatment period, the plant height and stem diameter of the woody plant species were re-measured and the leaf number of the graminoid plant species was re-counted. The plants were then harvested using the same method as for the pre-treatment harvest. To calculate the biomass growth of each plant, we subtracted the mean biomass of the five pre-treatment harvest plants for their species from their post-treatment measurement. To calculate the plant height (woody), stem diameter (woody) and leaf number (graminoid) growth of each plant, we subtracted their pre-treatment measurement from their post-treatment measurement. In addition to the growth metrics, the root-to-shoot ratio (R:S) of each plant was calculated.

Data analyses

Before conducting the data analyses, the data were checked to ensure they fulfilled the normality and equal variance assumptions of ANOVA. When these assumptions were violated, data were log-transformed.

To determine whether the assimilation rate, growth metrics and R:S of the study species differed between the watering × biostimulant treatment combinations, three-way ANOVAs were used. Watering treatment, biostimulant treatment and species were the fixed factors in the model. The ANOVAs were conducted for each growth form separately. When significant interactions were found, Tukey post-hoc analyses were performed to determine the treatment differences. Banksia robur was not included in the biomass growth and R:S analyses of the woody plant species because their cluster roots made root washing difficult, so an accurate measure of root biomass could not be obtained.

The data analyses were performed using R (v. 4.2.2, R Development Core Team 2022), with a significance level set at 0.05.

Results

Woody plant species growth metrics

There was a significant interaction between watering treatment and woody plant species for assimilation rate (F5,289=13.21, p < 0.001), with the well-watered plants having greater assimilation rates than the drought-stressed plants for all the species except C. australe (Fig. 1). There was no biostimulant effect on the assimilation rate of any of the woody plant species (F3,289=0.10, p = 0.960).

Fig. 1
figure 1

The assimilation rate for each watering treatment × woody plant species combination. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisks represent significant differences at p < 0.05

Full size image

There was a significant interaction between watering treatment and woody plant species for biomass growth (F4,240=3.04, p = 0.018), with the well-watered M. affine plants having greater biomass growth than the drought-stressed plants (Fig. 2). The biomass growth of the other five woody plant species did not differ between the watering treatments. There was no biostimulant effect on the biomass growth of any of the woody plant species (F3,240=0.79, p = 0.498).

Fig. 2
figure 2

The biomass growth for each watering treatment × woody plant species combination. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisk represents a significant difference at p < 0.05

Full size image

There was a significant interaction between watering treatment and woody plant species for plant height growth (F5,288=6.93, p < 0.001), with the well-watered S. smithii plants having greater plant height growth than the drought-stressed plants (Fig. 3a). The biomass growth of the other five woody plant species did not differ between the watering treatments. There was a significant biostimulant effect on the plant height growth of all of the woody plant species (F3,288=3.48, p = 0.016), with the seaweed biostimulant increasing growth more than the control treatment and protein hydrolysate biostimulant, but not the humic acid biostimulant (Fig. 3b).

Fig. 3
figure 3

The plant height growth for each (a) watering treatment × woody plant species combination and (b) biostimulant treatment. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Biostimulant treatment abbreviations: C = control, HS = humic substance, PH = protein hydrolysate, SW = seaweed. Each box represents the distribution of data within the interquartile range. The asterisk and letters represent significant differences at p < 0.05

Full size image

There was no significant interaction between watering treatment, biostimulant treatment and woody plant species for stem diameter growth. However, there was a watering treatment effect (F1,288=14.91, p < 0.001), with the well-watered plants having greater growth than the drought-stressed plants (Fig. 4). There was no significant biostimulant effect on the stem diameter growth of any of the woody plant species (F3,288=2.29, p = 0.079).

Fig. 4
figure 4

The stem diameter growth of the woody plant species for each watering treatment. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisk represents a significant difference at p < 0.05

Full size image

Graminoid plant species growth metrics

There was a significant interaction between watering treatment and biostimulant treatment for the assimilation rate of the graminoid plant species (F3,97=2,96, p = 0.036). The well-watered plants had greater assimilation rates than the drought-stressed plants for all the graminoid plant species (Fig. 5). Within the well-watered watering treatment, the control plants had greater assimilation rates than plants that received the seaweed biostimulant but not the humic acid and protein hydrolysate biostimulants.

Fig. 5
figure 5

The assimilation rate for each watering treatment × biostimulant treatment combination for the graminoid plant species. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The letters represent significant differences at p < 0.05

Full size image

There was a significant interaction between watering treatment and graminoid plant species for biomass growth (F2,144=8.20, p < 0.001), with the well-watered L. longifolia plants having greater biomass growth than the drought-stressed plants (Fig. 6). There was no significant biostimulant effect on the biomass growth of any of the graminoid plant species (F3,144=0.84, p = 0.476).

Fig. 6
figure 6

The biomass growth for each watering treatment × graminoid plant species combination. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisk represents a significant difference at p < 0.05

Full size image

There was no significant interaction between any of the factors for the leaf number growth of the graminoid plant species. However, there was a significant watering treatment effect (F1,144=8.78, p = 0.004), with the well-watered plants having greater growth than the drought-stressed plants (Fig. 7). There was no significant biostimulant effect on the leaf number growth of any of the graminoid plant species (F3,144=2.33, p = 0.080).

Fig. 7
figure 7

The leaf number growth of the graminoid plant species for each watering treatment. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisk represents a significant difference at p < 0.05

Full size image

Root-to-shoot ratio

There was no significant interaction between watering treatment, biostimulant treatment and species for the R:S of both growth forms. However, there was a watering treatment effect (woody: F1,240=9.08, p = 0.003; graminoid: F1,144=27.34, p < 0.001), with the drought-stressed plants having higher R:S than the well-watered plants (Fig. 8). There was no significant biostimulant effect on the R:S of both growth forms (woody: F1,240=0.49, p = 0.687; graminoid: F1,144=0.71, p = 0.546).

Fig. 8
figure 8

The root-to-shoot ratio of the (a) woody and (b) graminoid plant species for each watering treatment. Watering treatment abbreviations: DS = drought-stressed, WW = well-watered. Each box represents the distribution of data within the interquartile range. The asterisks represent significant differences at p < 0.05

Full size image

Discussion

We found that the stem diameter and leaf number (i.e. stem growth) of the woody and graminoid plant species respectively were significantly reduced by drought stress. Similar results have been reported in a range of urban plant species (Ugolini et al. 2012; Kagotani et al. 2016; Elansary 2017; Stratopoulos et al. 2019; Zhang et al. 2020; Dervishi et al. 2022). For example, Nitschke et al. (2017) found that the radial stem growth of English oak (Quercus robur), English elm (Ulmus procera) and Dutch elm (Platanus acerifolia) planted in Melbourne, Australia has been negatively impacted by historical drought events. Surprisingly, the reduction in shoot growth in the drought-stressed plants translated into an overall biomass growth reduction in only two of the study species (M. affine and L. longifolia). This is because these two species reallocated their resources from shoot to root biomass (i.e. higher R:S) to maximise their ability to obtain soil water. Although biomass growth was only minimally affected by drought stress, the significant reduction in assimilation rate amongst all the study species shows clearly that the plants were drought-stressed. This suggests that the plants were utilising their non-structural carbohydrate reserves to maintain their biomass growth while their assimilation rate was compromised (Hartmann and Trumbore 2016; Piper and Paula 2020). It is only when these non-structural carbohydrate reserves are diminished that a reduction in biomass growth would be observed (McDowell 2011). Therefore, we suggest that a longer-term experiment may have yielded more significant reductions in biomass growth amongst the drought-stressed plants.

Contrary to our hypotheses, we found very limited evidence that biostimulants improved plant growth or mitigated the impacts of drought stress on plant growth. In fact, the only benefit biostimulant application had on the plants was the seaweed biostimulant enhanced the plant height growth of the woody species compared to the control and protein hydrolysate biostimulant. Although this result is consistent with the findings of a recent meta-analysis that reported plant-based biostimulants (i.e. seaweed) are the most effective (Li et al. 2022), the general lack of responsiveness of our study species to biostimulant application supports the general trend reported in the urban forestry literature (Fraser and Percival 2003; Percival 2006; Banks and Percival 2012, 2014 see Ferrini and Nicese 2002 for exception).

There are a range of factors that can influence the effectiveness of biostimulants in promoting plant performance. The application method is important, with soil application reported to be more effective than foliar and seed application (Li et al. 2022). Although the contrary would be expected, given that foliar and seed application allows faster uptake of bioactive ingredients (Niu et al. 2021), the discrepancy between our findings and the biostimulant literature is not explained by the application method as we used soil application. Application frequency and dosage may also be important in plant responsiveness to biostimulants. In terms of frequency, a single application has been reported to be more effective than subsequent applications (Li et al. 2022). We carried out a single application, so application frequency does not explain the lack of responsiveness in our study species. Further, we followed the dosage instructions recommended by the biostimulant manufacturers, which suggests application dosage also does not explain our results. Finally, biostimulants are most effective in plants growing in suboptimal conditions such as low water availability (Li et al. 2022). The lack of interactive effects between the watering and biostimulant treatments on the performance of our study species suggests that the growth conditions also do not explain our results. Given that we seemingly used the ideal practices and conditions to yield a positive response in our study species to biostimulant application, yet observed only limited effect, raises questions about the efficacy of biostimulants in promoting plant performance.

To accurately gauge the potential benefit biostimulant application could have in Australian urban forestry, there are still efficacy issues that need to be resolved. Firstly, there is a clear publication bias in the biostimulant literature toward positive results (Li et al. 2022). Secondly, scientific field-based evidence on the efficacy of biostimulants is often not required for approval (Yakhin et al. 2017; Van Oosten et al. 2017). This is currently the case in Australia (Australian Pesticides and Veterinary Medicines Authority 2022), which means it is difficult for land managers (e.g. farmers and horticultural practitioners) to be certain about a biostimulant’s effectiveness. Considering these issues as well as the varying effectiveness of different application practices, it is likely that receiving beneficial outcomes from biostimulant application is highly context-dependent. In monoculture agricultural systems where environmental conditions are more uniform than other environments (e.g. urban, natural), biostimulant application may prove a useful practice as highlighted by the literature (Bulgari et al. 2015; Van Oosten et al. 2017; Li et al. 2022). However, in urban forestry, where there is a higher species richness and heterogenous environmental conditions, implementing effective biostimulant application practices may be challenging both in terms of logistics and cost. However, the first step in achieving this goal is to develop a framework that standardises how we measure the effectiveness of biostimulants, which will hopefully enable us to better understand the complex array of plant responses to biostimulants (Ricci et al. 2019).