Fire regimes are prevalent worldwide, and have been ascribed mostly to climate change, which results in rising temperatures and exacerbated droughts (Bowman et al., 2020; Bowd et al., 2022). Whitman et al. (2019) anticipated that the frequency and intensity of wildfires in various biomes would rise during the next 100 years. In South Africa, about 1.3 million ha of land are burnt annually (Russell-Smith et al., 2021), resulting in economic losses ranging between $123–215 million USD (Strydom and Savage, 2016). Consequently, a total of about 126, 000 fire incidences were recorded across South Africa in 2021, and fires mainly occurred in the Eastern Cape, North-West, and Mpumalanga provinces (Madondo et al., 2022). As a result, these incidences hamper the fulfilment of Sustainable Development Goals (SDG) 13 of Climate Action, which calls for immediate action to address climate change and its consequences, and SDG 15 of Life on Land, which aims to preserve the environment and save soils (United Nations, 2020).

Nevertheless, the extent of fire varies according to the type of biome. Among nine biomes found in South Africa, the savanna is the most fire-prone biome, constituting about 24.1% of the burnt area (Russell-Smith et al., 2021). It is imperative to note that fire is an important tool used in restructuring and managing savanna biomes due to their characterized mixture of grasses and scattered trees (Smith et al., 2013; Southworth et al., 2016). In this biome, fire is an important ecological factor for maintaining ecosystems and balancing competition among woody and herbaceous species (Ribeiro et al., 2019). Furley et al. (2008) deduced that savanna vegetation is typically fire-adapted and resilient. However, the increase in fire severity has resulted in the restructuring of vegetation and soil microbial communities. Thus, leading to long-lasting severe effects on ecosystem health and functions (de Groot et al., 2013).

Prescribed burning has been widely used to reduce the risk of severe wildfire occurrence (Fernández et al., 2013). Several studies have reported the effects of different prescribed burning frequencies on vegetation, vertebrate, and invertebrate species, as well as soil physical and chemical properties in the semi-arid savanna (Oluwole et al., 2008; Williams et al., 2012; Snyman, 2015; Holden et al., 2016; Madikana, 2022). For example, Madikana (2022) investigated the impact of various fire burning frequencies on the recovery of soil invertebrate communities in a semi-arid region of the Eastern Cape Province (South Africa) and observed a decrease in the abundance and diversity of earthworms, ants, and other ground-dwelling insect species. Another study conducted on savanna ecological zones in Ghana showed that anthropogenic fires increased soil pH, which had a positive impact on soil biological recovery (Amoako and Gambiza, 2019). Furthermore, it was demonstrated that fire critically altered soil physicochemical properties like changes in soil albedo and deterioration of nutrient pools (Magomani and van Tol, 2019). Fire can have a positive impact on soil nutrients P as burning transforms the organic P pool into orthophosphate, this is the only type of P accessible to biota, with P bioavailability being around the neutral pH range (Certain, 2005).

However, there is little knowledge concerning the impact of fire frequency on soil microorganisms (Fontúrbel et al., 2012; Certini et al., 2021). In all the soil microorganisms, fungi are the most susceptible to heat, and this could cause a change in ecosystem processes conducted by fungi (Dove et al., 2019). For instance, Vermeire et al. (2021) showed that species diversity was relatively lower in fungi than bacteria after various prescribed fire regimes. This is because microbes such as bacteria are relatively more tolerant of direct fire than fungi (Pulido-Chavez et al., 2023). However, fire causes a decrease in nutrient availability, such as carbon (C), nitrogen (N), and other organic substances like cellulose, which rapidly affect fungi (Adkins et al., 2020; Srikanthasamy et al., 2021). In terms of biomass and physiological activity, fungi are typically ranked as the most prevalent soil microorganisms (Kjoller and Struwe, 1982). They are abundant in the ecosystem and perform a variety of critical ecological tasks, such as nitrogen and carbon cycling and soil stabilization (Christensen, 1989; Boer et al., 2005; Janion-Scheepers et al., 2016). In addition to being decomposers, fungi are also saprobionts, mutualists, and parasites (Christensen, 1989; Challacombe et al., 2019). Fire regimes consequently affect nutrient cycling, and fungi play significant roles in nutrient cycles through nutrient transformations and enhancing plant-soil nutrient dynamics, mostly through symbiotic mycorrhizal fungi that enhance phosphorus absorption (Agbeshie et al., 2022).

Fire has direct effects on fungi through combustion and elevated temperatures, which cause mortalities (Dooley and Treseder, 2012; Nyamadzawo et al., 2013; Barreiro et al., 2016; Liu et al., 2023). Available evidence showed that savanna fires, which usually occur between 100°C and 700°C, are fatal to fungal communities, and their mortalities start happening as soon as the temperature reaches 60°C (Hansen et al., 2019). Indirect alteration of other soil properties makes it difficult for microbes to survive, which can happen through loss of habitat as well as food and energy (Singh et al., 2021). However, due to regular fire occurrence in savanna biomes, recurrent and frequent fires significantly shape the pre- and post-fire abundance of fungi in the soil (Hansen et al., 2019).

Considering the available evidence on how fire affects soil microbial activity, little is known about the effect of different prescribed fire frequencies on fungal diversity and community composition, particularly in the savanna rangeland. Also, the mechanisms that fungi use during their recovery process after fires require further research. In 1980, the University of Fort Hare in South Africa initiated a long-term trial aimed at evaluating the impact of different fire burning frequencies on various ecological indicators (Trollope, 1984). Since its establishment, the studies on this trial have been narrowly focused on the effect of fire on vegetation, with soil microorganisms such as fungi, which play a critical role in ecosystem function, being ignored. Therefore, the current study was conducted in a similar trial with the intention of assessing the effect of different fire burning frequencies on the diversity and taxonomic composition of soil fungal microbiota in a semi-arid savanna rangeland. It is hypothesized that soil microbial composition and diversity will be influenced by various fire frequencies.

The current study showed that fire increased soil fungi diversity, which was evident by higher diversity in burnt plots than unburnt (control) plots. The increase in fungi diversity and richness observed in this study after fire burning is attributed to various factors. The comparison of burnt treatment plots showed that frequently burnt treatment plots (annual, biennial, and triennial treatments) had higher diversity than the less frequently burnt treatment plots (quadrennial and sexennial treatments). This depicts that fungi diversity differs with the frequency of burning. The findings of this study were comparable to those made by Bowd et al. (2022), which reported high increases in fungi communities due to high fire frequency in the forest ecosystem.

The high diversity found in frequently burnt plots might be because repeated burning can create more resistance to fire and intensify the support of spore gemination (Hopkins et al., 2021). In frequently burnt plots, the influence can be reduced due to less vegetation recovery, such that when burning occurs, less heat is produced, which decreases its belowground dissemination. Conversely, in less frequently burnt plots, there is elevated vegetative recovery than in frequently burnt plots, and high heat amounts are produced, which increase fungi diversity in the soil (Semenova-Nelsen et al., 2019). Correspondingly, a previous study conducted on the same site demonstrated higher vegetative recovery on triennial, quadrennial, and sexennial burning frequencies when compared to frequently burnt treatments such as annual and biennial (Mopipi, 2012). Also, the conditions in less frequently burnt plots could be worsened by the indirect effects of fire. For instance, high heat loads result in large alterations of soil ecosystem features such as habitat loss, nutrient loss, and carbon loss, which provide shelter, food, and energy, and their recovery takes time (Alem et al., 2020).

In the current study, six taxa that were most altered by fire frequency included Mortierellomycota, Ascomycota, Basidiomycota, Chytridiomycota, Rozellomycota, and unknown. The largest representation of phyla was generally found in Ascomycota, which plays a critical role in litter decomposition, especially cellulose decomposition (Huffman and Madritch, 2018). In arid environments, Ascomycota fungi play a key role in the cycling of carbon and nitrogen (Challacombe et al., 2019). Fungi serves important roles in soil stability, plant biomass breakdown, and plant endophytic interactions. This large representation of this phyla has been reported in other fire frequency trials, including the study by Egidi et al. (2016), with a representation of 40% under grassland ecosystems, and Vázquez-Veloso et al. (2022), with a representation of 43% under Mediterranean ecosystems. Moreover, Bowd et al. (2022) also reported similar results, with a representation of 55% in the dry-sclerophyll forest ecosystem. This is because Ascomycota is the largest phylum in the fungi kingdom (Alem et al., 2020; Martín-Pinto et al., 2023), so it is usually expected to dominate in many studies. When the burnt plots were compared to the unburnt, the Ascomycota phyla showed the greatest resilience to fire because the abundance of its genus on all burnt plots increased, despite the fire frequency. Thus, on annual, biennial, triennial, and sexennial treatment, the Ascomycota phyla had the greatest representation of genus, which showed the utmost increase after fire burning. In addition to the Ascomycota phyla, the most common genus was Papulaspora, which increased more in annual, biennial, and sexennial burning.

The increase in Ascomycota phyla reflects its resistance to fire, especially with increased burning frequency, specifically in its genus Papulaspora. In the literature, a study conducted by Martín-Pinto et al. (2023) reported that post-fire burning, the altered soil environmental status results in the instant dominance of Pyrophilous (fire-loving) ascomycetous fungi, which ultimately results in Basidiomycota-Ascomycota formation. Salo et al. (2019) deduced that the Pyrophilous species in Ascomycota are highly attracted to burnt forests, which in this study could probably relate to the high numbers of Ascomycota in frequently burnt treatment plots. Pyrophilous species are attracted to burns. This is further facilitated by the substantial mineralizable matter presence post-fire burning, which brings into play the significance of organic matter influencing soil microbes (Bowd et al., 2022).

A study done in natural temperate grassland in Australia reported that the genus Dothideomycetes was mainly found under the phylum Ascomycota (Egidi et al., 2016). Another study in a dry Afromontane Forest in Ethiopia found Hypocreales, Pleosporales, and Chaetothyriales genus. In the savanna biomes, there is limited documentation on this aspect, which prompted the undertaking of the current study. The Ascomycota phyla are associated with carbon and nitrogen cycling, specifically in arid regions where they drive litter decomposition and plant endophytic interactions. In addition, it was reported that the Ascomycota fungi are associated with symbiotic associations as well as being saprotrophs or pathogens hosted in plant tissues (Challacombe et al., 2019).

The Coniochaeta species, which belong to the Ascomycota phylum, were also recorded in various burning treatments in the current study. Coniochaeta species can degrade lignocellulose in several woody substrates, including corn stover, sawdust, switchgrass, wheat straw, and coffee residues (van Heerden et al., 2011; Ravindran et al., 2012; Weber et al., 2015; de Lima Brossi et al., 2016). In addition, Coniochaeta species tend to have poor virulence on most hosts, and they usually colonize dead tissue or occupy already infected, injured, or dead plant tissues (Leonhardt et al., 2018).

Basidiomycota phyla was the second dominant after Ascomycota in the current study, and this dominance has also been reported in studies done in other ecosystems like dry-sclerophyll forest (Bowd et al., 2022), Mediterranean (Vázquez-Veloso et al., 2022), and pine savanna (Hopkins et al., 2021). Basidiomycota can also be Pyrophilous (same as Ascomycota), implying that they are capable of sporulating and forming colonies prior to fire subjection; therefore, such fungi sprout and become dominant soon after fire (Hopkins et al., 2021; Wang et al., 2023). Several classes under this phylum were higher under frequently burnt treatment plots (annual, biennial, and triennial treatments) than under less frequently burnt treatment plots (quadrennial and sexennial treatments). In addition, Basidiomycota classes such as Mycena, Papillotrema, Clavaria, and Lactimaria increased while Attractella and Ceratobasmidium decreased under these frequently burnt treatment plots. Basidiomycota are detritivores that play a critical role in ecosystem functioning. They contribute to carbon cycling through their nutrient absorption abilities during litter degradation (Taylor et al., 2014).

In this study, the Chytridiomycota phyla had less abundance in comparison to other phyla, which was evident by the presence of a few classes. This phylum (Chytridiomycota) is known saprophytically and is dominant in grassland environments. In these grassland ecosystems, it was found to be higher in burnt than unburnt plots (Egidi et al., 2016; Vázquez-Veloso et al., 2022). In the current study, the classes identified consist of Sonography lactis, which increased under annual and decreased under triennial burning, implying they are more tolerant of high than low frequent fire burning. Another class identified was C. gen. Incertae sedis, which increased under both triennial and quadrennial burning, implying that this class can be more tolerant in less frequently burnt soils. Chytridiomycota are known to make use of organic residues from fire burning in search of nutrients. Also, Chytridiomycota can increase when there is less competition with other fungi (Vázquez-Veloso et al., 2022). Glomus, a genus in the phylum Glomeromycota, was the most reduced when comparing biannual to no burning treatments. The phylum Glomeromycota contains an important fungus known as arbuscular mycorrhizal fungi (AMF) (Stürmer, 2012). The AMF form symbiotic relationships with the roots of many terrestrial plants, allowing them access to nutrients in the soil in exchange for C (Mpongwana et al., 2023). Longo et al. (2014) reported that fire occurrence negatively affected AMF spore communities in the Mountain Chaco Forest.

Mortierellomycota phyla were also less abundant in this study, and this was observed mostly on biennial and triennial burning treatments. The classes identified were Mortierella and Linnemania under biennial burning and Mortierellaceae_gen_Incertae_sedls under triennial burning. In addition, all these classes increased under both burning frequencies, which implied that this phylum could survive and increase under highly frequent fire regimes. This phylum has been identified in other fire studies done in dry sclerophyll forest (Bowd et al., 2022), subtropical monsoon (Wang et al., 2023), and boreal forest (Whitman et al., 2019). This phylum is known to inhabit several environments, such as the rhizosphere and plant tissues, where it mediates the carbon cycle and organic matter mineralization. In addition, Mortierellomycota are also plant-growth-promoting fungi (Muneer et al., 2021).

The current study showed that various fire frequencies had a positive impact on soil fungal abundance and diversity. Fire results in the alteration of the soil ecosystem, which facilitates the spore germination of numerous fungal species (Alem et al., 2020). If the fire is of low intensity, it may have little consequence on the fungi’s mortality (Egidi et al., 2016; Dove and Hart, 2017). Fungi communities make use of post-burning changes that increase abundance (Pressler et al., 2019; Vázquez-Veloso et al., 2022). For instance, from a biome perspective, savanna biomes in South Africa are characterized by a sparse population of trees rather than a high wood density responsible for high fire intensity. Another study has reported that after fire, ash deposits can be beneficial to the fungi and there can be less competition for resources with other species (Alem et al., 2020). The current study implies that, despite the devastation effects of fire frequency in savanna ecosystems, fire may increase soil fungi diversity and abundance, which can be significant in post-fire recovery due to the promotion of fungi, which play critical roles in soil functioning.

Generally, fungal metabarcoding studies that employ short-read high-throughput sequencing (HTS) technologies, like Ion Torrent, Illumina, and 454 Pyrosequencing, typically focus on the rDNA internal transcribed spacer regions (ITS1 or ITS2) (Lindahl et al., 2013; Furneaux et al., 2021; Cheng et al., 2023). In the current study, we used Ion Torrent sequencer which is a semiconductor-based next generational sequence (NGS) platform. Similarly to the Roche/454 system, the sensor’s pH change has poor linearity in proportion to the total number of nucleotides integrated in a single reaction cycle, decreasing its ability to identify homopolymer regions (Pu and Xiao, 2017; Cheng et al., 2023). In comparison to other platforms, the Illumina platform has the highest market share for sequencing tools. It can entirely address concerns in homopolymer sequencing, albeit there is a tendency for substitution errors in AT-rich and CG-rich sections (Dohm et al., 2008; Minoche et al., 2011; Cheng et al., 2023). In Illumina data, all sequence reads generated during a single experiment are the same length, however Ion Torrent reads vary. Furthermore, the latest generation of Illumina instruments can produce sequence reads from both ends of a fragment (“paired-end” reads), but Ion Torrent cannot (Brown et al., 2013). Unlike Illumina sequencing, multiple nucleotides may be incorporated during a single Ion Torrent sequencing cycle, and it is recognized that errors in quantitating the length of homopolymer repeats are common (Dickie, 2010; Wydro, 2022). As such, to better understand these discrepancies in sequence quality, further systematic comparisons between Ion Torrent and other HTS technologies are required, as have been done for various other platforms (Luo et al., 2012; Brown et al., 2013). Comparing these is crucial since different sequencing technologies differ in their capacity to sequence at different depths and lengths, as well as in terms of their quality profiles and possible biases (Furneaux et al., 2021).

In the current study, we noted that the quality of the sequences was not good as almost half of the samples failed the per base sequence quality, while the remainder passed with a warning. To improve the sequences, bbmap and DADA2 were used to remove the adapters and trim and filter quality. However, it is acknowledged that strict quality filtering could skew our assessment of the significance of particular fungal groups by eliminating them before the analysis. This study was limited to one geographical location by design; therefore, analyzing multiple rangelands within the same savanna biome in South Africa and elsewhere may provide a more complete picture of the effect of fire rates on soil fungal diversity and composition. Also, the distance between the burning treatments was 5 m, which could make it difficult to control the fire and as such the distance between treatments could be increased in future studies. It is necessary to point out that the techniques utilized in this study, (like all others), have inherent biases. As such, there is a potential for under representation of soil fungal diversity and strains that were beneath detected levels and were unable to develop under the fire frequency environments used. Therefore, it is critical to understand the benefits and drawbacks of employing next-generation sequence metabarcoding to detect and monitor significant functional groups on an ecological scale (Makiola et al., 2019).

The current study showed that different fire frequencies had an influence on soil fungi diversity and taxonomic composition in a semi-arid savanna rangeland. It was shown that the highest soil fungal microbial diversity was found in frequently burnt treatment plots (annual, biennial, and triennial treatments) than in less frequently burnt treatment plots (quadrennial and sexennial treatments). This could suggest fungi may survive more frequently occurring fires than less frequent ones. Regarding the taxa, Ascomycota and Basidiomycota were the phyla with the highest relative abundance followed by Mortierellomycota, Chytridiomycota, and Rozellomycota. Although there was heterogeneity in terms of how they were influenced by different fire frequencies, it was demonstrated that numerous phyla increased their communities with high fire frequencies over those with low fire frequencies. Although this is a long-term trial, the data in the current study was collected over one season, and it is recommended that future studies should collect data over a longer duration to assess the impact of prescribed fire frequencies on soil microorganisms over time. Such studies should also consider climatic conditions and environmental alterations in order to make informed decisions regarding conservation of fungi species in semi-arid savanna rangelands.