Response of weed flora dynamics in organic manuring and inorganic fertilisation in a smallholder crop-livestock system

Abstract

Fertiliser input in annual cropping systems can benefit crops and weeds. The effects of organic and inorganic fertilisers on weed species composition were evaluated across two agroecological zones in Eswatini: the Highveld and the Middleveld. Weed field surveys and weed seed bank studies were conducted over two cropping seasons, 2020/2021 and 2021/2022, in fields managed by smallholder farmers practising maize and cattle production. Fertiliser regimes were cattle manure only, cattle manure plus inorganic fertiliser, and inorganic fertiliser only, replicated six times in each region. The weed survey was conducted using three 25-m transects, and the weed seed bank was assessed at sampling depths of 0–10 and 10–20 cm. Results showed that cattle manure only and cattle manure plus inorganic fertiliser regimes had higher species richness and diversity than the inorganic fertiliser only regime. However, weed seed distribution did not differ between sampling depths or fertiliser regimes. Twelve weed species, including bermudagrass, crabgrass, and goosegrass, were key contributors to the differences in species composition between fertiliser regimes and agroecological zones, and were the most prevalent weeds in the seed bank. The study established that fertiliser type influences weed species composition and diversity. Proper cattle manure composting is essential to mitigate weed proliferation and enhance sustainable maize production. These findings demonstrate the need for integrated weed management and fertiliser strategies to improve crop yield and agricultural sustainability for smallholder farmers.

Introduction

Weed management is critical in smallholder crop-livestock systems due to limited farming inputs and varying management practices. Among the various strategies utilised, organic manure and inorganic fertiliser are key resources for soil fertility and crop productivity. Organic manure, which is generated from cattle for most smallholder crop-livestock farmers, improves soil fertility and provides essential nutrients for many field crops, including maize (Li et al., 2022; Maharjan et al., 2021; Materechera, 2010), but its nutrient composition is often insufficient to support optimal crop production (Little et al., 2021).

Uncured cattle manure, which many smallholder farmers use, can contain large amounts of viable weed seeds, and their passage through the digestive system can stimulate germination and increase weed diversity and biomass when the manure is applied to cropping fields (Mkhabela, 2006; Mkhabela and Materechera, 2003; Materechera and Modiakgotla, 2006; Pleasant and Schlather, 1994). The viability of seeds passing through cattle’s digestive systems is highly influenced by hardiness (Harrington et al., 2011), dormancy, and time spent in the digestive system (Rahimi et al., 2016). Kasirivu et al. (2011) reported that seed viability of Lamb’s quarters (Chenopodium album L.) and smooth pigweed (Amaranthus hybridus L.) was most dominant with manure application and that barnyardgrass [Echinochloa crus-gali (L.) Beav.] exhibited dormancy, which in turn increased survival of passage through cattle.

Conversely, inorganic fertilisers provide a precise nutrient supply when applied at recommended rates, supporting crop development, while having differing impacts on weed species communities. For example, nitrogen (N) fertiliser influences the density of various weed species, while phosphorus (P) fertiliser affects the occurrence of weed species (Travlos et al., 2018). Fertilisers containing N and P promote the growth of nitrophilous and phosphophilous weeds (Lehoczky et al., 2014; Nadeem et al., 2018; Ņečajeva et al., 2015; Tang et al., 2014), but there is a paucity of studies on the specific effect of potassium (K). Nonetheless, NPK fertilisers increase the potency of weed propagules and enhance rapid weed growth (Nwagwu et al., 2015).

Even so, there is limited information on the influence of cattle manure and inorganic fertilisers on weed composition in crop-livestock systems. Moreover, most research on the value of cattle manure as a fertiliser has been conducted in areas with growth conditions that differ significantly from those in Sub-Saharan Africa’s crop-livestock systems (Lithourgidis et al., 2007; Ozlu et al., 2019). Thus, the study aimed to assess the effects of cattle manure and inorganic fertiliser on the floristic and seed bank composition in smallholder crop-livestock systems in Eswatini. Diversity indices, including the Shannon-Wiener, Simpson, and Pielou index, were employed to characterise species composition. The study also utilised the importance value index (IVI) to identify weeds that significantly contributed to the seed bank. The hypothesis tested was that applying cattle manure might increase weed species composition in fields where it is applied, thereby exacerbating challenges in weed control.

Materials and methods

Study site

The study was performed in small-scale farmers’ fields of two agroecological regions of Eswatini, namely, the Highveld (Mankayane; 26° 40′ S, 31° 3′ E, elevation 900–1800 m.a.s.l) and the Middleveld (Luve; 26° 20′ S, 31° 28′ E, elevation 400–800 m.a.s.l) (Figure 1). During the summer, maize is the main crop grown under rainfed conditions; therefore, the weed field survey and weed seed bank studies were conducted between October and May for two cropping seasons (i.e. 2020/2021 and 2021/2022). In the 2020/2021 seasons, the Highveld recorded total rainfall amounts of 904 and 780 mm in 2021/2022, accompanied by mean temperatures of 20.9°C and 18.4°C, respectively. Conversely, the Middleveld experienced 808 mm of rainfall in 2020/2021 and 505 mm in 2021/2022, with mean temperatures of 23.2 and 21.7°C, respectively. Soil at Mankayane was a sandy clay loam with a clay content of 28 %, while at Luve was a sandy loam with 19 % clay. The average soil pH was measured at 4.3 for Mankayane and 4.7 for Luve.

Figure 1. Map showing the two agroecological zones of Eswatini, with triangles indicating the sampled/surveyed fields in the (a) Highveld (Mankayane) and (b) Middleveld (Luve).

Cropping field and farming practices

Stratified sampling technique (Müller-Dombois and Ellenberg, 1974) was used to select major maize-growing agroecological regions, specifically the Highveld and Middleveld of Eswatini, and a Regional Development Area (RDAs) within each region. Within a 2 km radius of each selected RDA, eighteen farmers with a history of maize production spanning five years or more were identified. These farmers were categorised into three fertiliser regimes: manure only (n = 6), manure plus inorganic fertiliser (n = 6), and inorganic fertiliser only (n = 6), resulting in three fertiliser regimes with six replicates each (Table 1).

Table 1. Fertiliser regime in the studied maize-growing agroecological zones of Eswatini ^a

^a NPK, Nitrogen Phosphorus Potassium; LAN, Limestone Ammonium Nitrate.

For land preparation, farmers utilised a mouldboard plough followed by a disc harrow. In the fertiliser regimes where manure was applied, farmers broadcast the manure before ploughing at an average rate of 9 t/ha (N = 33 kg, P = 55 kg, K = 70 kg) only in the first year, with manure applied every two cropping seasons. The manure was sourced from the kraal, where cattle are confined nightly and accumulate throughout the year. At the start of the cropping season, the manure is collected and transported to the fields for broadcasting as the primary fertiliser. During the winter dry season, cattle feed on maize residues from the previous harvest, which further contributes to nutrient recycling. It is important to note that this method does not ensure complete manure composting, potentially containing viable seeds (Mncube, 2023).

Maize variety SC 719 (Seed-Co®, Zimbabwe) was planted at a seeding rate of 25 kg/ha, at inter- and intra-row spacing of 0.9 and 0.25 m, respectively. Basal fertiliser [N: P: K, 2:3:2 (37%)] was applied at a planting rate of 500 kg/ha at Mankayane and 400 kg/ha at Luve. Four weeks after crop emergence (WACE), Limestone Ammonium Nitrate [LAN (28%)] top dressing fertiliser was applied at 200 kg/ha at Mankayane and 150 kg/ha at Luve. Weed management was through early post-emergence (EPOST) herbicide application at the maize four-leaf stage using cyanazine (Bladex® 900 g/kg, AgNova Technologies Pty Ltd, Hamilton, Australia) at the recommended dosage rate (4 L/ha). A PB-20 or PB-16 Knapsack sprayer (Jun Chong Co., Ltd., Kluang, Johor, Malaysia) was calibrated to reach and maintain 300 kPa pressure, with a spray volume of 200 L/ha. Cyanazine was chosen for application because it is readily available and commonly used for weed management in maize, especially considering the limited variety of herbicides available in the Eswatini market (Dlamini et al., 2016).

Soil seed bank sampling

The characterisation of the soil weed seed bank species was conducted before land preparation (before manure broadcasting and planting) over two cropping seasons. Plots measuring 20 m × 20 m were marked in each field. Soil cores were collected from each plot at 2 m intervals in a zigzag pattern using a 5 cm diameter auger in 2020 (October 10^th–18^th) and 2021 (October 15th–22^nd). Sampling was performed at depths of 0–10 cm and 10–20 cm, with ten samples per depth and site. These samples were combined to create composite mixtures per site and stored in paper bags at room temperature until spread in plastic trays. In 2020 (October 25^th) and 2021 (October 28^th), plastic trays measuring 30 cm × 30 cm x 10 cm were filled with 4 kg of sterilised dry sand to achieve a soil depth of 3 cm in the tray, then samples from each field and depth were thinly and evenly spread to a depth of 2 cm. The base of the trays was perforated to allow drainage. Trays were laid in a randomised complete block design (RCBD) under natural light in a screen house, with six replicates. Throughout the 20-week maize growth period, the soil in the trays was watered daily to stimulate germination. Every two weeks, the soil (excluding sand) in each tray was turned using a hand fork to stimulate seed germination of photoblastic weeds. Germinated weed seedlings were identified, counted, and removed.

Weed sampling survey

Weed surveys were conducted in 2021 (February 26^th–April 4^th) and 2022 (March 2^nd–April 14^th). Fields were surveyed at the maize kernel dough stage (i.e., R4 growth stage). Post-emergence herbicide application occurred two months before the R4 stage, influencing weed populations and contributing to the weed seed bank and flora of subsequent cropping seasons. To assess the structure and abundance of weed vegetation in maize fields, three transects were marked in each field: two along the field edges (vertically) and one at the field centre (horizontally). Each transect was 25 m long and comprised five 5 m × 1 m quadrats at 1 m intervals. This sampling design was chosen to accommodate field sizes ranging from 50 m × 50 m to 100 m × 60 m.

Established plants of broadleaved weeds with a minimum of 10 cm in height, grass weeds of 15 cm in height, and stoloniferous weeds with a minimum length of 10 cm were sampled during the study period. Weed species were identified using screening methods from the Royal Botanic Gardens (Kew) and online databases, such as POWO (2022) and CABI (2022). Weed species density and composition were assessed in the weed seed bank and maize field survey studies. Species in ≤5% of the sampling quadrats were considered rare, while those in more than 25% were defined as common.

Data processing

The number of individual weed species that germinated in the weed seed bank experiment was converted to weed density/m². The absolute and relative values of frequency, density, and abundance for each species were used to analyse the structure of the weed seed banks (Curtis and McIntosh, 1950) for each fertiliser regime, agroecological zone (study area), and season. Absolute weed density was determined as the total number of weed species per sampled area (Equation 1), while relative density was the absolute density of a weed species over the sum of all absolute densities (Equation 2).

(1) $${\rm{ Absolute \,weed\, density\! = \!{Total\;number\;of\;targeted\;species}}/\\ {{Total\;sampled\;area}}}$$

(2) $${\rm{{Relative\; density =[Absolute\; density\;of\;weed\;species}}/ \\{{Sum\;of\;all\;absolute\;densities] \times 100}}}$$

Absolute frequency was the total number of sampling units (fertiliser regime replicates) with the targeted species over the total number of sampling units (Equation 3).

(3) $${ \small{\rm{Absolute\; frequency} = {Number\;of\;sampling\;units\;with\;targeted\;species}/}\\{{\small\rm{Total\;number\;of\;sampling\;units}}}}$$

The relative frequency was the absolute frequency of each species over the sum of all frequencies (Equation 4).

(4) $$\eqalign{& {\rm{Relative\; frequency = [Species\;absolute\;frequency}}/\cr& {\rm{Sum\;of\;all\;absolute\;frequencies] \times100}}}$$

Absolute dominance was the total number of individuals per species over the total number of sampling units containing that species (Equation 5), while relative dominance was the absolute dominance of a species over the sum of all dominances (Equation 6).

(5) $${\eqalign{& {\rm {Absolute\; dominance = {Total\;number\;of\;individuals\;of\;a\;species}}}/\cr& {\rm{Total\;number\;of\;sampling\;units\;containing\;that\;species}}}}$$

(6) $${ {\rm{Relative\; dominance = [Absolute\;dominance\;of\;a\;species}}/\\{\\\rm{Sum\;of\;all\;absolute\;dominance]\times 100}}}$$

The IVI was calculated as the sum of relative frequency, relative density, and relative dominance (Equation 7). The index ranges from 0 to 300 per sample, and the weed species with the highest value is considered the most important.

(7) $$\small{\rm {Importance\; value\; index = \sum {\matrix{{{\rm{Relative}}\;{\rm{density}}\;{\rm{ + }}} \cr {\;{\rm{Relative}}\;{\rm{frequency}}\;{\rm{ + }}\;{\rm{Relative}}\;{\rm{dominance}}} \cr } }}} $$

The diversity indices were used to determine species composition (Legendre and Legendre, 1998) between the fertiliser regimes, agroecological zones, and sampling seasons in the seed bank and maize field survey. Species richness (S) was the total number of species per sample (Equation 8).

(8) $${\rm{S = }}\sum\limits_{{{i = 1}}}^{{s}} {pi^0} $$

where s equals the number of species, p _i equals the ratio of individuals of species i, and superscript 0 is the insensitivity to species frequencies.

Shannon-Weiner (H’) diversity index was the relative abundance of each species and the species richness between the fertiliser regimes, seasons, and agroecological zones (Equation 9).

(9) $$H' = \sum\nolimits_{{\rm{i = 1}}}^{\rm{s}} {{p_i}\,{\rm{in}}\,{p_i}} $$

Gini-Simpson (D) diversity index was used to measure the degree of dominance of weed species within the fertiliser regimes, seasons, and agroecological zones (Equation 10).

(10) $${\rm{D = 1}} - \sum\limits_{i = 1}^s {[{{\rm{n}}_i}({\rm{n}} - 1)} /{\rm{N (N}} - {\rm{1)]}}$$

where n is the number of individuals of one species, N = the total number of all individuals.

Pielou (J’) evenness index was used to measure the degree to which the abundances of different weed species in a fertiliser regime, season, or agroecological zone are similar (Equation 11).

(11) $$J' = H'/Log(S)$$

Data analyses

A Generalised Linear Model (GLM), followed by post hoc pairwise analysis with Least Square Difference (LSD) correction for multiple pairwise comparisons, was used to determine whether there was a difference in the diversity indices of fertiliser regimes, seasons, agroecological zones, and sampling depths for the soil seed bank and their interactions. Analyses were conducted in Statistical Package for Social Science (IBM SPSS Statistics version 28), and statistical significance was accepted at p < 0.05.

Weed survey data were square-root transformed to balance the weight of the contribution of the most abundant and rare species. To calculate the similarity index between fertiliser regimes, the transformed data were subjected to the Bray-Curtis similarity index (Anderson et al., 2008). Weed species composition (the number and kinds of species) between fertiliser regimes, agroecological zones, and seasons was compared using Permutational Analysis of Variance (PERMANOVA) with 999 random permutations. This analysis was further supported by Principal Coordinate Analysis (PCoA) to visualise the compositional differences across factors. Similarity Percentage (SIMPER; Clarke, 1993) analysis was used to calculate the contribution of each species to the dissimilarities between the species compositions of fertiliser regimes, agroecological zones, and seasons. Explored species were those that contributed a dissimilarity of ≥2.5%. Analyses were conducted in PRIMER + PERMANOVA (Plymouth Marine Laboratory Routines in Multivariate Ecological Research, version 6: Plymouth Marine Laboratory, UK).

Results

Cropping field weed survey

The cropping field weed survey study identified 66 weed species from 18 families (Supplementary Table S1). Six weed species were found exclusively in Mankayane (Highveld), while three were unique to Luve (Middleveld). Southern cutgrass (Leersia hexandra Swartz.), cheese weed (Malva parviflora L.), witch grass (Panicum capillare L.), and yellow foxtail [Setaria pumila (Pior.) Roem. & Schult.] were only observed in the first season. In contrast, dwarf papyrus sedge (Cyperus haspan L.), and dodder (Cuscuta australis R.Br.) were only observed in the second season. Redroot pigweed (Amaranthus retroflexus L.), spiny pigweed (Amaranthus spinosus L.), balloon vine (Cardiospermum halicacabum L.), jimson weed (Datura stramonium L.), common purslane (Portulaca oleracea L.), linden leaf (Salvia tiliifolia Vahl.), and black nightshade (Solanum nigrum L.) were found to be associated with manure application.

Regarding the ecological community, 41 rare (R) and 25 common (C) species were recorded (Supplementary Table S1). The observed species exhibited various life forms, including 41 annual broadleaved weeds (AB), one annual leafless weed (AL), six perennial broadleaved weeds (PB), nine annual grass weeds (AG), five perennial grass weeds (PG), and six perennial sedge weeds (PS). Weed species distribution showed that 26 species were native (N) to the agroecological zones, while 40 were introduced (I).

The composition of weed species from the weed survey differed significantly in the interaction between fertiliser regimes and agroecological zones (Pseudo F-value = 3.024, P <= 0.001), and between fertiliser regimes, agroecological zones, and seasons. The PCoA on the species composition for the interaction between fertiliser regimes and agroecological zones revealed three main components with eigenvalues greater than one, which explained 45.3 % of the variation in weed species composition (Figure 2). In addition, PCO1 and PCO2 explained 33.6 % of the total variation. The samples clustered according to the agroecological zone regardless of the fertiliser regime, with minimal overlap (Figure 2).

Figure 2. Principal Coordinate Analysis (PCoA) comparing weed species composition between fertiliser regimes and study areas based on Bray-Curtis similarity matrix.

Weed species richness (S) was influenced by the interaction between fertiliser regimes and agroecological zones (F _{(72, 2)} 9.273, P < 0.01). In the Middleveld, species richness was significantly higher in the manure only and manure plus inorganic fertiliser regimes and lower in the inorganic fertiliser only regime (Figure 3). In the Highveld, the manure only, manure plus inorganic fertiliser, and inorganic fertiliser only regimes were lower than those of the Middleveld (Figure 3).

Figure 3. Mean (±SD) weed species richness (S) between fertiliser regimes and study areas. The different letters indicate post hoc pairwise differences.

Furthermore, the Shannon-Wiener (H’) diversity index differed significantly between the fertiliser regimes (F _{(72, 2)} 9.608, P < 0.01); weed species diversity in the manure only regime was greater than that of the inorganic fertiliser only regime (Figure 4a). Also, H’ differed significantly between agroecological zones (F _{(72, 1)} 33.7, P < 0.001); weed species diversity in the Middleveld was greater than in the Highveld (Figure 4b).

Figure 4. Mean (±SD) weed species diversity (H’) between fertiliser regimes (a) and study areas (b). The different letters indicate post hoc pairwise differences.

When considering the degree of weed species dominance within fertiliser regimes, seasons, and agroecological zones, the Gini-Simpson (D) index differed significantly between agroecological zones only (F _{(72, 1)} 16.8, P < 0.001). Dominance in the Middleveld was lower compared to the Highveld (Figure 5a). In terms of the degree to which the abundances of different weed species in a fertiliser regime, season, or agroecological zone are similar, Pielou’s (J’) evenness also differed significantly between the agroecological zones (F _{(72, 1)} 11.3, P < 0.001) and was greater in the Middleveld compared to the Highveld (Figure 5b).

Figure 5. Mean (±SD) weed species dominance (D) between study areas (a). Mean (±SD) weed species evenness (J’) between study areas (b). The different letters indicate post hoc pairwise differences.

The Similarity Percentage (SIMPER) analysis showed uniform within average species composition similarities in the Highveld, with values of 47, 48, and 47 % for the manure only, manure plus inorganic fertiliser, and inorganic fertiliser only regimes, respectively (Table 2). In the Middleveld, the SIMPER analysis revealed within average species composition similarity of 67 % in the manure only regime, which was higher than the manure plus inorganic fertiliser (62 %) and inorganic fertiliser only (59 %) regimes. The average species composition dissimilarities in the Highveld were lower between manure only and manure plus inorganic fertiliser (54 %) regimes, compared to manure only and inorganic fertiliser only (60 %), manure plus inorganic fertiliser and inorganic fertiliser only (59 %) regimes (Table 2). Similarly, in the Middleveld, the average species composition dissimilarities were lower between manure only and manure plus inorganic fertiliser (37 %), compared to manure only and inorganic fertiliser only (42 %), manure plus inorganic fertiliser and fertiliser only (44 %).

Table 2. Within fertiliser regime similarity and between fertiliser regimes pairwise dissimilarities of weed species as determined by SIMPER ^a

^a M, manure only; F, fertiliser only; M + F, manure plus fertiliser.

SIMPER identified 12 dominant discriminating weed species contributing an abundance of ≥2.5 % to the weed species composition dissimilarities observed between the fertiliser regimes and agroecological zones (Table 3). These weed species contributed to the significant differences in species composition, richness, diversity, dominance, and evenness and the similarities and dissimilarities observed between the fertiliser regimes and agroecological zones. Also, five weed species were the most important and discriminating in the seed bank of the different fertiliser regimes (Supplementary Table S1). Three (C. dactylon, D. sanguinalis and E. indica) were abundant in the seed bank and field survey.

Table 3. Similarity Percent (SIMPER) analysis on weed species contributing more than 2.5% between fertiliser regimes and agroecological zones ^a

^a M, manure only; F, fertiliser only; M + F, manure plus fertiliser.

Weed seed bank

The weed seed bank study revealed 49 weed species in the first season and 45 in the second. Weed species were from 19 families (Supplementary Table S1). Forty-four weed species were common between the seed bank and the cropping field survey, while five were present in the seed bank only. The dominance of weed species in the seed bank was summarised using the importance value index (IVI) (Supplementary Table S1). During the first season, under the manure only regime, E. indica was dominant (>38) in the Highveld, while C. dactylon was dominant (>31) in the Middleveld. Under the inorganic fertiliser only regime, C. dactylon was dominant (>68) in the Highveld, whereas G. parviflora was dominant in the Middleveld. The manure plus inorganic fertiliser regime showed the dominance of E. indica (>78) in the Highveld, while D. sanguinalis (>39) was dominant in the Middleveld. During the second season, the manure only regime showed dominance of S. tiliifolia (>39) in the Highveld, while C. dactylon was dominant (>26) in the Middleveld. The manure plus inorganic fertiliser regime showed dominance of C. dactylon (>31) in the Highveld, while D. sanguinalis was dominant (>39) in the Middleveld. The inorganic fertiliser only regime showed dominance of D. sanguinalis (>56) in the Highveld, while G. parviflora was dominant (>26) in the Middleveld.

In terms of weed seed density, there were no significant differences between sampling depths of 0–10 cm and 10–20 cm (F _{(72, 1)} 0.792, P > 0.05) and fertiliser regimes (F _{(72, 2)} 4.903, P > 0.05). However, weed seed density was influenced by the agroecological zones (F _{(72, 1)} 20.1, P < 0.001). The density was lower in the Highveld compared to the Middleveld (Figure 6a). Additionally, weed seed density differed across sampling seasons (F _{(72, 1)} 17.1, P < 0.001); a higher density was observed in the second season compared to the first season (Figure 6b).

Figure 6. Mean (±SD) weed seed density between study areas (a) and seasons (b). The different letters indicate post hoc pairwise differences.

Weed seed bank species richness (S) demonstrated a significant interaction between the fertiliser regimes and agroecological zones (F _{(72, 2)} 9.291, P < 0.01). Species richness (S) was lower in the inorganic fertiliser only regime in the Highveld, compared to the Middleveld (Fig. 7a). Seed bank diversity (H’) showed a significant interaction between the fertiliser regimes and agroecological zones (F _{(72, 2)} 19.4, P < 0.001). Species diversity was lower in the inorganic fertiliser only regime in the Highveld compared to the Middleveld. Moreover, species diversity was greater in the manure only regime compared to the manure plus inorganic fertiliser regime in both agroecological zones (Figure 7b).

Figure 7. Mean (±SD) weed species richness (S) (a) and Shannon-Wiener (H’) diversity (b) for the interaction between fertiliser regimes and study areas. The different letters indicate post hoc pairwise differences.

Discussion

The study aimed to distinguish weed communities in both the field and seed bank between maize fields fertilised with cattle manure and inorganic fertiliser in two agroecological regions of Eswatini. The higher weed species richness observed in the manure only and manure plus fertiliser regimes in the Middleveld compared to the Highveld could be due to the region’s warmer temperatures, which can enhance weed seed germination and survival. The warmer temperatures in the Middleveld could have affected soil temperature, influencing seed germination and the rate of non-dormant seeds (Klupczyńska and Pawłowski, 2021; Peng et al., 2021). These results are consistent with previous research (e.g., Haj Sghaier et al., 2022; Koger et al., 2004; Steinhaus et al., 2002), which showed that weed seed germination increases linearly with temperature up to an optimal point.

The reduced species richness in the fertiliser only regime could be attributed to the ability of certain weed species to quickly respond to fertility, allowing them to grow and dominate (Jiang et al., 2014; Rotchés-Ribalta et al., 2017). Continuous fertiliser application often encourages adaptable weed species, enhancing their competitive abilities and contributing to the seed bank (Blackshaw et al., 2003; Little et al., 2021). The results corroborate those of De Cauwer et al. (2011) and Jiang et al. (2018), who reported a significant reduction in weed species richness and diversity with inorganic fertiliser applications.

Inorganic fertilisers, unlike uncured manure, effectively reduce weed species’ richness and diversity, supported by herbicide applications such as cyanazine. This herbicide reduced the abundance of susceptible broadleaved weed species, favouring grasses (Hyvönen and Salonen, 2002). Indeed, nine annual grasses, six perennial grasses, and six sedges, which were unlikely to be affected by cyanazine application, were recorded in the maize cropping fields. Most grasses (e.g., D. sanguinalis, E. indica, and C. dactylon) and sedges [e.g., yellow nutsedge (Cyperus esculentus L.) and purple nutsedge (Cyperus rotundus L.)] exhibited high abundances, indicating insufficient control by cyanazine. Additionally, the fertiliser only regime had the lowest weed biomass and the highest maize yield compared to the manure only and manure plus inorganic fertiliser regimes in both sampling seasons and study areas (Mncube et al., 2024).

Comparisons of weed species assemblages in the Highveld revealed low similarities within fertiliser regimes and high dissimilarities between fertiliser regimes. Conversely, the Middleveld showed high similarities within fertiliser regimes and low dissimilarities between fertiliser regimes. This was supported by the significant differences in species evenness (J’) across the agroecological zones, indicating heterogeneous weed communities in the Highveld and homogeneous ones in the Middleveld. The variability suggests greater species composition within Highveld cropping fields and higher percentages of rare species, contributing to overall diversity. Weed dispersals from neighbouring habitats, such as forests and grasslands, enhance weed biodiversity (Gabriel et al., 2005, 2006; Roschewitz et al., 2005). In the present study, sampling was within a 2 km radius, and as the distance between fields increased, diversity is likely to have increased as well. The findings align with those of Alignier and Petit (2012), who observed significant dissimilarities in the weed community in fields separated by ≥450 m.

On the other hand, the observed homogeneous weed species in cropping fields indicate an evenly distributed assemblage, suggesting reduced diversity, influenced by proximity between fields, which encourages propagule movement (Gaba et al., 2010). Additionally, crop competition within fields can reduce weed abundance and richness, leading to a uniform weed assemblage (Berquer et al., 2021). Further, consistent cropping practices result in a more stable weed community structure, limiting diversity.

This study identified 12 weed species in the field weed survey that contributed to significant differences in species composition, richness, diversity, dominance, and evenness, as well as similarities and dissimilarities observed between the fertiliser regimes and study areas. Additionally, five weed species were most important in the seed bank amongst the different fertiliser regimes; three (Cynodon dactylon, Digitaria sanguinalis, and Eleusine indica) were abundant in both the seed bank and field weed survey. This indicated that the presence of these weed species was not influenced by fertiliser regime or agroecological zone, suggesting that they might be generalists capable of thriving in different field conditions, such as a wide range of soils, light availability, and cooler or warmer temperatures (De Cauwer et al., 2021; Fried et al., 2020; Sivarajan and Pradeep, 1994).

According to Romaneckas et al. (2021), field cultivation plays a key role in dispersing weed seeds by burying them at different depths. However, this study’s findings show that weed seed density was consistent and did not differ between the sampling depths of 0–10 and 10–20 cm. This consistency could be attributed to the tillage methods used by farmers. Specifically, mouldboard ploughing in this study may have led to the even distribution of weed seeds across soil layers. The results agree with those of Clements et al. (1996) and Feledyn-Szewczyk et al. (2020), who found that mouldboard ploughing and conventional tillage produce a uniform distribution of weed seeds throughout the soil profile, unlike minimum tillage practices. This even distribution across soil layers indicates a large weed seed bank capable of supporting high seed survival rates. As a result, managing different weed species may require specific strategies to reduce seed production and prevent continuous addition to the soil seed bank.

The seed bank is influenced by many factors, including soil texture (Benvenuti and Mazzoncini, 2021). Therefore, the manure and inorganic fertiliser application history of more than five years could have influenced the physical properties. As such, the particle assembly in manured fields could have promoted the soil’s suitability for weed seed survival. Moreover, seed germination in aggregate soils could be accelerated by the availability of moisture, air, warmth, and light, which breaks weed seed dormancy (Benvenuti, 2003; Long et al., 2015). Hence, weed seeds germinating in aggregate soils may not require increased energy reserves to germinate and can divert this energy toward seedling establishment and growth (Benvenuti and Mazzoncini, 2019). Consequently, manure applied fields would have elevated seed diversity and richness, with early-emerging weed seeds that could be competitive during the early stages of crop growth. Additionally, manure reduces soil compaction, enhances soil aggregate stability, and improves porosity and water infiltration, which may increase seed richness and diversity by stabilising particles that enable weed seeds to survive at various soil depths (Rayne and Aula, 2020).

Furthermore, observations from both the seed bank and field survey indicated that the relationship between above- and below-ground vegetation can be attributed to the continuous addition of manure, which is rich in weed seeds. This could have increased the weed species richness and diversity in the soil (Materechera and Modiakgotla, 2006; Mkhabela, 2006; Mkhabela and Materechera, 2003; Pleasant and Schlather, 1994; Rupende et al., 1998), which supports the hypothesis that manure applied fields would have a higher weed species composition than inorganic fertiliser applied fields. Therefore, in addition to the EPOST herbicide used by the farmers in the study, weed management techniques aiming at reducing weed emergence and competitiveness can be considered to minimise weed seed richness and diversity. Also, farmers may need to consider composting cattle manure to kill weed seeds before broadcasting it in the field.

Conclusion

The study demonstrated that fertiliser regimes have limited short-term effects on weed seed distribution in soil but significantly influence weed species richness and diversity. The manure only and manure plus inorganic fertiliser regimes exhibited higher species richness, with native, adaptable species thriving in disturbed areas; therefore, farmers need to understand the behaviour of these weeds to implement cropping and management practices that reduce their success. Successful weed management of both homogeneous and heterogeneous communities will require techniques that reduce weed densities and minimise the invasion of new species, which current weed management strategies cannot control. A long-term, ecologically sustainable weed management programme that is easy for farmers to understand is essential for developing effective strategies.