Pathogenic Species of Botryosphaeriaceae Involved in Tree Dieback in an Urban Forest Affected by Climate Change

TL;DR

Urban forests are suffering from tree dieback due to climate change and thinning practices, which increase the spread of harmful fungi.

Contribution

The study identifies specific pathogenic Botryosphaeriaceae species linked to tree dieback in urban forests affected by thinning and climate stress.

Findings

Thinning increased the incidence of Botryosphaeriaceae pathogens in declining forest plots.

Neofusicoccum parvum and Botryosphaeria dothidea increased tenfold and fivefold, respectively, in thinned subplots.

Thinning altered microclimate and microbial balance, promoting harmful fungi proliferation.

Abstract

Urban forests are highly valued for the multiple benefits they provide to city dwellers. The strategic provision of ecosystem services by these forests is threatened by climate change, warming conditions being responsible for heat waves and chronic droughts that inflict stress and mortality on trees. A three-year study (2011–2013) conducted at Parco Nord Milano (PNM) (Milano, Italy) assessed the impact of thinning interventions on the dynamics of fungal pathogens in declining forest plots. Symptomatic trees of the genera Alnus, Acer, Fraxinus, Platanus, Quercus and Ulmus, exhibited in thinned subplot pronounced decline/dieback, exhibiting symptoms like microphyllia, leaf yellowing, leaf shedding, sunken cankers, shoot wilting and branch dieback. Comparative analyses between the thinned and unthinned subplots revealed a significantly higher incidence of pathogens in the thinned one. Five species of Botryosphaeriaceae, namely Botryosphaeria dothidea, Diplodia corticola, Diplodia seriata, Dothiorella omnivora and Neofusicoccum parvum, were consistently isolated from tissues of declining hosts. There is evidence that thinning altered plot-level microclimate conditions and microbial equilibrium, favoring the proliferation of latent, pathogenic Botryosphaeriaceae. In fact, during the study period, the presence of N. parvum increased tenfold and that of B. dothidea fivefold in thinned subplot. Conversely, in unthinned subplot, the same pathogenic taxa maintained stable proportions. These results demonstrate that thinning altered ecological balances increasing tree susceptibility to harmful, cosmopolitan botryosphaeriaceous fungi. Our findings challenge assumptions about thinning as a universally beneficial practice, emphasizing the need for silvicultural strategies that take into account host and pathogen ecology and the microclimatic resilience of forest stands. This study emphasizes the importance of adaptive management in urban forestry to mitigate the unintended ecological consequences of climate change.

Figures

Tables

Table 1. Morphological characteristics and GenBank accession numbers of Botryosphaeriaceae species isolated in this study.

Species	Colony Morphology (PDA)	Conidia (Shape and Septation)	Conidial Size (µm)	Pycnidia/Conidiomata
Botryosphaeria dothideaPX780779	Colonies initially olivaceous, becoming dark grey to black with age; moderately dense mycelium with smooth margins	Fusoid to ovoid, unicellular, hyaline, rarely septate before germination	17.0–22.0 (mean 19.5) × 7.7–10.0 (mean 8.1)	Globose, solitary, ostiolate, producing hyaline conidia
Neofusicoccum parvumPX780728	Fast-growing; abundant aerial mycelium, initially white, turning grey to black with age; reverse similarly darkening	Fusiform, unicellular, hyaline	17.5–25.0 (mean 20.8) × 7.5–10.0 (mean 7.9)	Dark, globose, ostiolate, produced on PDA
Diplodia corticolaPX780741	Dark, stromatic conidiomata, immersed to partially erumpent, multiloculate and ostiolate	Oblong to cylindrical, thick-walled, hyaline, aseptate, rarely septate with age	23.5–46.0 (mean 29.9) × 9.0–18.5 (mean 13.6)	Stromatic, multiloculate, ostiolate
Diplodia seriataPX780777	White to grey mycelium, becoming dark grey to olive with age	Ovoid to oblong, mostly aseptate, occasionally 1-septate at maturity; hyaline to brown	17.6–22.4 (mean 20.0) × 8.1–11.2 (mean 9.6)	Dark, globose, ostiolate
Dothiorella omnivoraPX780735	Grey to dark olive-green mycelium, moderate growth	Oblong to ovoid, aseptate, becoming 1-septate at maturity; hyaline turning brown	20.8–24.0 (mean 22.4) × 11.2–14.4 (mean 12.8)	Dark, globose, ostiolate

Table 2. Environmental parameters recorded at PNM over a three-year period in the two subplots, unthinned (2Au) and thinned (2At).

Parameters	Unthinned Subplot (2Au)			Thinned Subplot (2At)
	2011	2012	2013	2011	2012	2013
Average Temperature (°C)	14.6	14.4	13.7	15.9	15.8	14.9
Maximum temperature (°C)	31.6	32.0	29.9	32.7	33.2	31.8
Average temperature of April and May (°C)	18.4	15.3	14.6	20.1	17.5	17.0
Rainfall (mm)	773.7	986.5	1167.9	773.7	986.5	1167.9
Light intensity (Lx)	4.000	3.500	3.000	16.500	18.000	17.500
Relative humidity (%)	75.64	79.85	75.85	52.65	59.18	58.56

Full text

1. Introduction

A number of forests have shown, in recent decades, signs of decline/dieback on a global scale [1,2]. Climate-driven changes, harbingers of physiological plant impairment and other disturbances (e.g., disease and pest outbreaks), have been recognized as major drivers of poor growth, dieback and mortality of trees in global forests [3,4].

The phenomenon appears increasingly exacerbated in urban forests, which are less resilient than forests growing in natural environments [5,6]. These forests exhibit heightened vulnerability mainly due to their anthropogenic origins. In fact, except for a few sporadic exceptions, in which remnants of pre-urban vegetation were incorporated into the urban fabric, urban forests are generally artificial and, as such, they lack ecological integrity [7]. Many urban manmade forests are a mosaic of native tree species intermixed with non-native ones, planted to ameliorate the aesthetics, but disregarding ecological considerations (e.g., the possible allelopathic interactions between species or unsuitable soil conditions). These urban forests, overly simplified in composition and structure, are exposed in the “urban jungle” to several stress factors, the most common being water deficiency, low moisture, excess heat, soil compaction (e.g., by intensive trampling) and pollution load [8]. These abiotic and anthropogenic stressors can induce tree dieback and mortality also in natural forests [9], but they may turn devastating, causing massive tree deaths, in low-resilience urban forests.

The inherent predisposition of urban forests to natural and man-induced injuries is normally skillfully handled by urban foresters who wisely manage the damage, so that the decline/dieback is not noticed by the citizens [10]. Indeed, urban forests having been conceived not for timber production but to offer several social services (e.g., landscaping, recreational, and psychological benefits) to the citizenship [11], intensive management is performed to make their deterioration less noticeable. For instance, understory management, the pruning of desiccated branches, phytosanitary thinning (cutting off dying or dead trees attacked by parasites), and removing hazardous trees are priority interventions in urban forests, so that urbanites can safely continue to frequent them. Timely and intensive management thus often masks the true extent of urban forest decline/dieback [12]. Without this special care and maintenance, their deterioration would be much more apparent, which could compromise the multiple functions and services for which they had been created.

The degradation of urban forests, common to many large cities, appears more and more pronounced in climate change hotspot zones. One such hotspot is the Mediterranean basin [13]. In this area, climate change is strongly altering weather regimes, with warmer temperatures, heat waves, water deficits, prolonged droughts during the growing season, and more frequent and intense weather events (e.g., windstorms, rainstorms and floods); these are having far-reaching impacts on the health of the trees in urban forests. Drought weakens trees and predisposes them to pathogen attacks; windstorms cause cracks and ruptures of twigs and branches, thereby facilitating infection by decaying agents. Hail, often associated with rainstorms, produces diffuse bark lesions which become preferential entry points for opportunistic pathogens; flooding spreads harmful “water moulds”, such as the destructive Phytophthora species [14,15,16]. Environmental change is thus a precursor to tree damage and attacks by an array of infectious agents which, either abruptly or gradually depending on each specific pathogen’s survival strategy and virulence, lead to tree death [17,18].

The Parco Nord Milano (PNM) is an artificial urban park in the city of Milano, one of the largest European cities. The park extends for over 600 hectares between the northern outskirts of Milano and five neighboring municipalities. PNM includes multifunctional green spaces (forests, grasslands, wetlands and river corridors) created over formerly agricultural and industrial lands (centuries-old factories that have now almost completely disappeared as a result of de-industrialization). Having arisen in the suburb of a metropolitan area, PNM also incorporates urban built-up areas and transport infrastructures that, over time, have connected the northern outskirts of Milano to its hinterland with no zoning plan. The rushed creation of this green space, which was intended to prevent the overbuilding of residual lands, meant that plots were rapidly reclaimed, greened and equipped with infrastructures for public use. This also entailed that the principles of urban ecology were not a priority in its planning [19]. For this reason, PNM today presents a heterogeneity of soils and wooded plots composed of maladapted tree assemblages.

Since PNM is situated in a vast metropolitan area within the Mediterranean climate change hotspot zone, it was attractive for testing hypotheses regarding the impact of climate-driven changes on the conditions of its wooded areas. Specifically, a phytopathological investigation was conducted to verify which biotic agents were possibly involved in the etiology of widespread tree mortality in some forest plots (Figure 1).

This paper reports the results of a three-year research study at PNM that aimed to (a) identify the causative agent(s) responsible for the decline/dieback symptoms observed on some trees in the park; (b) test whether a relationship existed between temperature/rainfall patterns and the abundance of some fungal tree pathogens; (c) assess if trees’ physiological impairment caused by climate constraints increased the occurrence of fungal tree pathogens; and (d) determine if, and to what extent, thinning, performed for both silvicultural and phytosanitary purposes, favoured pathogenic fungal taxa.

2. Materials and Methods

2.1. The Study Site

This research was conducted at Parco Nord Milano (PNM; 45°27′47″ N, 09°11′16″ E, elevation 121 m a.s.l.), a peri-urban park in Lombardy, northern Italy. The tree populations of the park exhibit heterogeneous configurations, including isolated trees, open meadows, clustered groups, linear alignments, multi-row strips, and dense forest patches. The investigation focused on a closed-canopy forest block (Plot 2A) containing two distinct subplots 2At and 2Au (45.538293, 9.212198 and 45.541051, 9.213550) subjected to different thinning management regimes (Figure 2), as defined in PNM’s long-term thematic maps (established since the park’s inception in 1983).

Subplots 2At (thinned) and 2Au (unthinned) comprised approximately 80 trees (diameter at breast height, DBH ≥ 10 cm) each, dominated by Alnus cordata (Loisel.) Duby., Acer platanoides L., Fraxinus angustifolia Vahl., F. excelsior L., F. ornus L., Platanus hispanica Mill. ex Münchh., Quercus cerris L., Q. robur L., Q. rubra L., and Ulmus pumila L. These tree species were the most abundant in terms of population composition and structure, accounting for over 90% of plant biomass, according to the Park’s forest inventory data. As regards their origin, A. platanoides and all species of Fraxinus and Quercus (with the exception of the exotic Q. rubra, native to the eastern and central regions of the United States and Canada [20]) are native to Italy, while F. ornus and Q. cerris are typical of central-southern Italy and the Apennines. A. cordata also grows in Lombardy but as a cultivated and ornamental species in parks and avenues, outside its natural range, which is southern Italy (Campania and Calabria) [21]. P. hispanica is a hybrid between the European plane tree (Platanus orientalis L.) and the American plane tree (Platanus occidentalis L.) [22]. U. pumila (Siberian elm) is an exotic species, native to East Asia, particularly Siberia and northern China [23].

The experimental design was conceived to assess the combined effects of thinning and drought-related stress under field conditions, rather than to disentangle their individual contributions. Subplot 2At underwent sequential silvicultural treatments, including a light thinning during dormancy (2008/2009), followed by a moderate thinning in 2010/2011. An extreme drought event in summer 2011, characterized by more than five consecutive months with temperatures exceeding 40 °C, triggered widespread canopy desiccation and required the removal of severely affected trees in February–March 2012. A subsequent medium-to-heavy thinning was applied in 2013 to mitigate residual stress. In contrast, subplot 2Au remained unmanaged throughout the study period, serving as a baseline for evaluating thinning-related effects. This experimental framework allowed a comparative assessment of forest structural dynamics and resilience under contrasting management regimes within a Mediterranean–Continental climate [24].

2.2. Fungal Isolation and Identification

Samplings were carried out once per year over three consecutive years (from April 2011 to May 2013), during the same seasonal period, on the same selected trees. At each sampling event, four current-year twigs (2–4 cm in diameter) were collected from the crown of each tree (one from each cardinal point) at a height of approximately 2.50 m from the ground. Each year, a total of 300 twigs per subplot were collected from the available trees, including both symptomatic and asymptomatic individuals. Asymptomatic plants were defined as those showing no visible disease symptoms at the time of sampling. Overall, this sampling design resulted in a total of 1800 twig samples, corresponding to approximately 9000 fragments, of which 4500 fragments were obtained from thinned subplots and 4500 fragments from unthinned subplots. Samples were placed in moist polyethylene bags and stored at 4 °C during transport and until use. The samples were always processed the day after collection in our laboratory in Florence.

Samples were initially examined for fungal structures (e.g., pycnidia, ascomata, basidiomata) using a Leica Wild M8 stereoscope (Leica Microsystems, Heerbrugg, Switzerland). Surface sterilization was performed by immersing tissue fragments in 3% sodium hypochlorite (NaClO) for 1 min, followed by rinsing with sterile distilled water. After air-drying on sterile absorbent paper, the outer bark was aseptically removed using a scalpel. Five tissue fragments (4–5 mm^2^) were placed on 2% potato dextrose agar (PDA) (Liofilchem Srl, Roseto degli Abruzzi, Italy) supplemented with 250 mg L^−1^ ampicillin (Sigma-Aldrich, St. Louis, MO, USA) and 10 mg L^−1^ rifampicin (Sigma-Aldrich, St. Louis, MO, USA) to inhibit bacterial growth. Plates were incubated in the dark at 24 ± 1 °C for 7–14 days; in the case of asymptomatic samples showing no initial microbial growth, incubation was prolonged and plates were periodically inspected for the appearance of microbial colonies. Emerging fungal colonies were purified using single-hypha and single-conidium isolation techniques, followed by subculturing onto fresh potato dextrose agar (PDA) to obtain pure cultures.

2.3. Morphological Identification

A preliminary screening was carried out to group isolates that showed differences in colony morphology (morphotypes). Morphotype identification was performed according to colony morphology (colour, margin, texture, compactness, surface topography and growth rate) and micromorphological characteristics (Table 1). To induce sporulation, isolates were transferred to water agar supplemented with autoclaved pine needles [25] and incubated at room temperature under UV light. After 14 days, pycnidia were excised and mounted in 100% lactic acid for microscopic examination using a Zeiss optical microscope (ZEISS, Jena, Germany). Conidial dimensions (length × width) were measured for 100 conidia per morphotype at ×400 magnification. A strong presence of Botryosphaeriaceae fungi was found to colonise most of the tissues examined, with most of the resulting morphotypes belonging to five predominant species. Isolates were preserved in the mycological culture collection of the Plant Pathology Section of the DAGRI Department, University of Florence.

2.4. Molecular Identification

To confirm morphological identification, pure cultures of botryosphaeriaceous fungal taxa were obtained from hyphal tips taken aseptically under a dissecting microscope (Leica Microsystems, Heerbrugg, Switzerland) and grown for 7 days on Malt extract agar (MEA) (Liofilchem Srl, Roseto degli Abruzzi, Italy) at 20 °C. Fungal DNA was then extracted from lyophilized mycelium ground under liquid N_2_ to a fine powder, as in Moricca et al. [26]. The rDNA ITS region was PCR-amplified using the universal primers ITS1 and ITS4 [27]. Amplification was performed in a reaction mixture (25 µL volumes) containing: 0.2 µM of each primer; 2.5 µL of 10× Taq DNA polymerase buffer (10 mM Tris HCl pH 8.3, 1.5 mM MgCl_2_, 50 mM KCl and 0.1 mg gelatin, Promega, Madison, WI, USA); 200 µM each of dATP, dCTP, dGTP, and dTTP (Promega, Madison, WI, USA); 10 ng of template DNA, and 0.5 units of Taq DNA polymerase (Promega, Madison, WI, USA). Thermocycler settings were 3 min for the initial denaturation at 95 °C, followed by 35 cycles of 30 s at 95 °C, 30 s at 55 °C, and 1 min at 72 °C; with a 5 min final extension at 72 °C [28]. A negative control with all reagents except DNA was included to verify that the reactions were free of contamination. PCR products were separated by electrophoresis through 1% (v/v^−1^) agarose gel (Sigma-Aldrich, St. Louis, MO, USA) containing 0.5 µg/L ethidium bromide (Sigma-Aldrich, St. Louis, MO, USA) and visualized under a UV transilluminator. Amplicons were prepared for sequencing using an Exo-SAP-IT enzymatic purification system according to manufacturer’s instructions (USB Corporation, Cleveland, OH, USA). ITS regions were sequenced in both directions with the PCR primers ITS1 and ITS4 used for amplification by the CIBIACI University service (Florence, Italy).

Fungal identity was determined by comparing the obtained consensus sequences with those deposited in NCBI using BLAST+ 2.17.0 (Basic Local Alignment Search Tool; http://www.ncbi.nlm.nih.gov/BLAST, accessed 28 September 2023) [29]. The generated sequences were submitted and deposited in NCBI GenBank. The sequences were aligned with sequences of closely and distantly related ex-type fungal taxa using ClustalX v. 1.83 [30] with the following parameters: pairwise alignment (gap opening = 10, gap extension = 0.1) and multiple sequence alignment (gap opening = 10, gap extension = 0.2, transition weight = 0.5, divergence sequence delay = 25%). Phylogenetic analyses were conducted using the Maximum Likelihood (ML) method implemented in MEGA X version 10.1.8, with all alignment gaps included. The optimal nucleotide substitution model was automatically selected by the software.

2.5. Isolation Frequency Assessment

The isolation frequency of Botryosphaeriaceae was calculated using the formula:

$$[eqn]$$

where F represents the frequency of Botryosphaeriaceae, NBot is the number of fragments from which Botryosphaeriaceae species were isolated, and NTot is the total number of wood fragments yielding fungal isolates [31].

The comparison between thinned and unthinned subplots was based on the average fungal assemblages isolated from symptomatic and asymptomatic trees of ten species: Alnus cordata, Acer platanoides, Fraxinus angustifolia, F. excelsior, F. ornus, Platanus hispanica, Quercus cerris, Q. robur, Q. rubra, and Ulmus pumila.

Statistical analyses were performed using IBM SPSS Statistics (version 29). Isolation frequency (IF) data were aggregated at the subplot level and expressed as mean values ± standard deviation. Prior to statistical analyses, data normality for each group was assessed using the Shapiro–Wilk test (p > 0.05). Differences between treatments (thinned vs. unthinned subplots) in terms of the percentage incidence of fungi belonging to the Botryosphaeriaceae family were evaluated using Student’s t-test for independent samples, with the level of significance set at p ≤ 0.05. In addition, comparisons of incidence between thinned and unthinned subplots over the study period were conducted separately for each year.

2.6. Microclimatic Parameters of the Subplots

Light intensity was measured using a photometer (Vianello Photovolt Model 200 L, Milan, Italy) positioned at the center of each subplot at a height of 1 m. Additionally, average temperature (°C) and precipitation data (mm) were collected from the PNM weather station over the three years. To explore the relationship between temperature and relative humidity data from the PNM weather station and those from the investigated subplots, a PCE 555 humidity/temperature meter (PCE Instruments, Lucca, Italy) was placed at the same location/height as the photometer. The thermometer was positioned in the shade. Climatic data are showed in Table 2.

2.7. Dendrometric Parameters of the Subplots

To measure forest density, in order to assess how crowded or open the two subplots were, the basal area, i.e., the cross-sectional area of each trunk, measured at breast height [DBH—Diameter at Breast Height, usually 1.3 m (4.5 ft) above the ground] was calculated in each subplot. With reference to a forest stand/plot, the basal area is the sum of the basal areas of all the trees present in it.

3. Results

Species identification was based on the analysis of 124 fungal colonies, integrating molecular characteristics (ITS sequences) with morphological characteristics (macro and micro morphology). Colony phenotypes and conidial morphology and size were summarized in Table 1. In addition, the isolates were compared with reference strains from our collection previously characterized using multilocus sequence analyses. This combined approach led to the identification of five members of the Botryosphaeriaceae family: Botryosphaeria dothidea, Neofusicoccum parvum, Diplodia corticola, D. seriata, and Dothiorella omnivora (Table 1).

DNA sequencing confirmed the identification of the five Botryosphaeriaceae species, showing 100% sequence identity with corresponding isolates deposited in the GenBank database. The isolates clustered into well-supported clades (ML bootstrap values > 95%) together with ex-type culture sequences (Figure 3).

In the first of the three sampling years, no significant differences were found between the thinned and the unthinned subplots. In the thinned subplot, the analyzed material showed 72.8% sterile fragments, 5.4% B. dothidea, 1.9% N. parvum, 4.5% D. corticola, 2.6% D. seriata, and 12.8% Do. omnivora. In the unthinned subplot, the results were 79% sterile fragments, 7.9% B. dothidea, 3% D. corticola, and 10.1% Do. omnivora. In the second year, results in the thinned subplot showed a reduction in sterile fragment percentage (42.1%) and an increase in pathogen presence: 15.1% B. dothidea, 11.8% N. parvum, 11.3% D. corticola, 7.1% D. seriata, and 13.2% Do. omnivora. In the unthinned subplot, the recorded values were: 70.1% sterile fragments, 6.9% B. dothidea, 8.3% N. parvum, 2.6% D. seriata, 4.3% D. corticola, and 7.8% Do. omnivora. In the third year, the thinned subplot showed a further increase in pathogen frequency: 21.5% sterile fragments, 24.7% B. dothidea, 22.1% N. parvum, 11.2% D. corticola, 7% D. seriata, and 13.5% Do. omnivora. In the unthinned subplot, the results remained more stable: 71.6% sterile fragments, 7.4% B. dothidea, 6.2% N. parvum, 4.2% D. seriata, 4.4% D. corticola, and 6.2% Do. omnivore (Figure 4).

For each of the three years, t-test analysis revealed a significant difference in the isolation frequency (IF) of Botryosphaeriaceae between thinned and unthinned subplots (p < 0.01; Figure 5). The species most frequently isolated from the thinned subplot, compared to the unthinned subplot, were D. corticola, D. seriata and Do. omnivora.

Meteorological data analysis revealed that the thinned subplot recorded higher temperatures (ranging from 1.1 °C to 1.99 °C higher) and lower humidity levels (reductions between 17.29% and 22.99%) compared to the unthinned subplot (Table 2). The measurement of the basal area proved to be of considerable practical importance, as it revealed how much of the ground surface was occupied by tree trunks in each subplot. According to data from the PNM forestry office, the basal area of the thinned subplot decreased from 3.79 m^2^ ha^−1^ in 2006 to 1.82 m^2^ ha^−1^ in 2012, while the basal area of the unthinned subplot decreased from 4.78 m^2^ ha^−1^ in 2006 to 4.45 m^2^ ha^−1^ in 2012. Although both subplots experienced a reduction in basal area due to abiotic and biotic factors (such as drought and pathogenic endophytes), thinning further influenced temperature, humidity, and light intensity.

Average and maximum temperatures were consistently higher in the thinned subplot than in the unthinned subplot across all years (Table 2). In particular, average temperatures in the thinned (2At) subplot exceeded those of the unthinned (2Au) by approximately 1.2–1.5 °C, while maximum temperatures were up to 1.6 °C higher. Spring temperatures (April–May), a critical period for host susceptibility, also showed higher values in the thinned (2At) subplot in all three years. At the interannual scale, April and May 2011 were characterized by higher temperatures (16.9 °C and 20.0 °C, respectively) compared to temperatures of the same months in 2012 (12.5 °C and 18.1 °C) and 2013 (13.4 °C and 15.8 °C). Rainfall values were identical for thinned (2At) and unthinned (2Au) subplots, as both were derived from the same local meteorological station and therefore reflected site-scale climatic conditions rather than treatment effects. At the interannual scale, precipitation was markedly lower in April and May 2011 (8.1 mm and 61.5 mm, respectively) compared to precipitation of the same months in 2012 (172.2 mm and 111.0 mm) and 2013 (156.0 mm and 177.0 mm) (Figure 6). The summer drought of 2013 exacerbated tree stress, leading to an increase in the IF of N. parvum, B. dothidea, D. corticola, and Do. omnivora. In the thinned (2At) subplot, 70% of the trees exhibited typical dieback and cankers associated with Botryosphaeriaceae infection [32].

4. Discussion

The findings of this study highlight a fundamental link between human activities (forest management), climatic stressors, and fungal pathogenicity in driving tree mortality within PNM. The increased isolation frequency of Botryosphaeriaceae fungi, particularly Botryosphaeria dothidea and Neofusicoccum parvum, in the thinned subplot aligns with growing evidence that forest management practices, when superimposed on environmental extremes, can amplify tree vulnerability to opportunistic pathogens [33]. These dynamics underscore the fragility of tree-endophyte symbioses under compounded stress, where latent fungal endophytes transition to virulent pathogens, accelerating tree decline [34]. The role of recurrent drought and heatwaves, such as the 2011 event, appears pivotal in this process. Prolonged water scarcity and thermal extremes likely depleted tree carbon reserves, inducing carbon starvation and weakening defensive responses, a phenomenon consistent with the carbon starvation hypothesis [34,35]. This physiological impairment may have facilitated fungal exploitation, as Botryosphaeriaceae are known to exploit stress-compromised hosts [36,37]. Botryosphaeriaceae are widely recognized as important pathogens of woody hosts, causing symptoms such as branch and crown dieback, stem cankers, and bark necroses [38,39]. A distinctive feature of many species within this family is their ability to persist as latent pathogens or endophytes and to express pathogenicity primarily when hosts are weakened by biotic or abiotic stress factors [32]. This ecological behavior provides a useful framework for interpreting the patterns observed in this study, where differences in isolation frequency among Botryosphaeriaceae species were associated with contrasting microclimatic and stand conditions.

Notably, the broad host range of B. dothidea, as indicated by the occurrence of identical haplotypes across multiple tree species, suggests a potential for cross-species transmission, which may partly account for the widespread mortality observed within the PNM [40,41]. This plasticity reflects patterns observed in other stressed ecosystems, where generalist pathogens thrive in low-diversity stands with reduced ecological redundancy [42]. It is therefore important to carefully evaluate the potential role of thinning as a catalyst for pathogen proliferation [43]. While thinning is often implemented to reduce competition and enhance stand resilience, the physical wounds generated during this process likely provided entry points for fungal colonization [44,45]. Botryosphaeriaceae are well-documented to exploit both natural openings and mechanical injuries, and the timing of thinning activities—coinciding with periods of climatic stress—may have exacerbated host susceptibility [46,47]. Post-thinning microclimatic shifts, such as increased solar radiation and reduced humidity, could further amplify water stress, particularly during heatwaves, creating a feedback loop that favors pathogen establishment [48,49]. These observations align with studies in Mediterranean and semi-arid ecosystems, where managed forests exhibit higher pathogen loads due to disrupted canopy buffering and heightened abiotic stress [50]. The loss of keystone species and reduced tree diversity in PNM likely compounded these effects, diminishing the ecosystem’s capacity to resist pathogen-driven destabilization [51,52].

The ecological and management implications of these findings are profound. In urban-adjacent forests like PNM, where anthropogenic and climatic pressures converge, traditional management practices may require re-evaluation to avoid unintended consequences. Adjusting thinning intensity, timing, or spatial patterning to minimize synergies with climatic extremes could mitigate pathogen facilitation. Furthermore, preserving biodiversity and keystone species may enhance ecosystem resilience by maintaining functional redundancy and stabilizing microbial interactions. Monitoring fungal communities following disturbances could serve as an early indicator of incipient decline, enabling proactive interventions. However, this study’s three-year scope and single paired-plot design limit causal inference, necessitating caution in generalizing findings. Long-term, replicated studies integrating soil health, insect dynamics, and fungal population genetics would be critical to disentangle the roles of biotic and abiotic drivers. Experimental approaches, such as controlled stress trials paired with pathogen inoculation, could further elucidate thresholds for virulence activation and host susceptibility.

Although total rainfall was identical between subplots, thinning is known to alter local microclimate (e.g., soil moisture, temperature and evapotranspiration) and epidemiological conditions (e.g., host physiological impairment and the amount of spores intercepted by individual trees during the spore fall). These factors may play a key role in shaping host susceptibility and pathogen dynamics.

In this study, members of the Botryosphaeriaceae, including B. dothidea, N. parvum, and D. seriata, were among the most frequently isolated fungal taxa in the PNM stands, and their isolation frequency increased in the thinned (2At) subplot, suggesting a strong response of these fungi to changes in stand structure and microclimatic conditions [53]. For years, in fact, the woody material derived from thinnings in PNM, even thinnings made for phytosanitary purposes, was chipped and then spread along the park’s avenues and paths (this is a common practice in many contexts, especially in urban green areas). This procedure was discontinued when we demonstrated that B. dothidea occurred on wood chips and remained viable for several months, surviving even at temperatures of 40 °C for 24 h. Wood chips thus served at PNM as an inoculum reservoir of the pathogen for a long time, allowing the microorganism to increase its inoculum pressure from year to year. Among other things, it is easy to imagine how this mass of inoculum, present on the wood chips scattered on the ground (often also on trails used by pedestrians and various vehicles), was moreover exposed to any potential vector [54,55].

Results indicate that thinning significantly altered the subplot’s microclimatic and microbiological equilibrium, leading to substantial increases in pathogenic fungi. Specifically, N. parvum exhibited a tenfold increase compared to the first year, while B. dothidea quintupled in prevalence. Concurrently, a marked reduction in sterile fragments suggested a decline in beneficial endophytic communities, facilitating opportunistic colonization. Botryosphaeriaceae display heightened aggressiveness under hydrothermal stress. In trees mechanically wounded by thinning operations, combined with elevated exposure to abiotic stressors (e.g., hail), the entry of pathogens was further amplified [56]. Pruning practices, often conducted without tool sterilization or wound protection, exacerbated infection risks [57,58,59].

Thinning drastically modifies stand microclimate, influencing multiple parameters: leaf area index (LAI), transpiration rates, soil water availability, temperature, humidity, and light regimes. Initially, thinning reduces LAI, diminishing canopy radiation interception capacity. Subsequently, LAI recovers to pre-thinning levels due to increased light intensity but becomes concentrated on fewer trees [56]. Elevated foliar mass per tree in thinned stands increases available substrates for foliar and shoot pathogens. Furthermore, as tree canopies act as natural spore traps [57], greater foliar mass per tree enhances spore interception during pathogen dispersal events. Thus, higher inoculum loads on canopies correlate with increased infection and colonization probabilities by airborne fungal pathogens [58,59]. Reduced tree density also lowers inter- and intraspecific competition for light and nutrients, theoretically reducing individual physiological stress [60]. Smaller basal area in thinned subplot, with lower density and ostensibly improved growing conditions, was expected to enhance tree vigour and reduce susceptibility to biotic stressors, including pathogens [61]. Paradoxically, this subplot exhibited the highest frequency of pathogenic Botryosphaeriaceae isolation. Conversely, in dense, unthinned stands, factors such as increased shading, humidity, tissue proximity, and nutrient limitation were anticipated to heighten fungal susceptibility. These findings suggest that stand density may influence endophytic pathogen infection rates and abundance in opposing directions. Dynamics typical of undisturbed or moderately disturbed forest ecosystems, where thinning enhances tree growth, vigour, and defences, may be therefore overturned in low-resilience artificial stands, such as urban forests heavily impacted by warming conditions.

The conventional assumption that reduced stand density improves tree vitality and lowers pathogenic fungal infection does not apply to chronically degraded urban forests. In such systems, trees subjected to environmental stress and anthropogenic disturbance often face depleted energy reserves, a condition that may negate the benefits of management practices [62,63,64]. At PNM, severe drought coupled with recent record-high temperatures depleted tree carbon reserves, inducing carbon starvation and elevating vulnerability to pathogenic endophytes and mortality [65]. However, management practices likely contributed to this outcome. Improper thinning operations—specifically, dispersing infected wood chips within the park, creating unprotected wounds, and using unsterilized pruning equipment—may have exacerbated pathogen spread [65]. These results highlight the importance of assessing how management strategies can influence forest responses under conditions of increasing environmental stress. Although longer-term monitoring could allow for a more accurate disentanglement of the individual effects of climate variability and forest management practices, the present study provides valuable information on the responses of artificial forest stands to pathogens in the short and medium term.

5. Conclusions

Urban forests are a fundamental resource for increasingly crowded metropolises, improving the urban environment and contributing to the well-being of citizens [66]. Understanding how best to preserve urban forests and the varied services they provide has today become a priority. Despite its importance, the impact of forest management practices, especially thinning, on urban forests, and more generally on stands with prevalently or exclusively tourist-recreational functions, has received scant attention. Moreover, studies on the influence of thinning on the fungal community harbored in urban forests are even more scarce, or even non-existent. Here we demonstrated that the harsh environmental conditions (mainly drought and heat stress) can not only frustrate management efforts (e.g., thinning) in urban forests, but even select for the most pathogenic component of the fungal endophytic community [67,68]. Climate change thus is driving the abundance and composition of the fungal endophytic assemblage in urban forests. This study is meant to inform and support foresters, park managers and restoration ecologists involved in the management or the repair of urban forest ecosystems.