Seven soil samples were collected from different durum wheat (Triticum turgidum var. L. durum) fields near Fez, Morocco (Latitude (dd): 34.03715, Longitude (dd): 4.998). To obtain a baseline of local soil degradation, a previously farmed area (now abandoned and uncultivated due to erosion and land degradation) in Blad Sadar was also sampled (DS, Table 1 ). Regional mean annual temperature and potential evapotranspiration between 1970 and 2011 at the Fez-Sais station was 16.9°C and 863 mm per year, respectively (35). Clay-loam is the dominant soil texture. Despite relatively poor moisture retention, light reddish siliceous soil (locally named “Hamri”) is typically used for wheat production in the area (36). Wheat fields in the region often contain large amounts of stones and gravel (37). Soil degradation is widespread due to dry conditions and intensive land use, accompanied by frequent ploughing and overgrazing. Average annual rain-fed wheat yield in the Fez region between 1989–2014 was 1.27 t ha^-1 (https://www.yieldgap.org/accessed on 19 November 2022), with 180 growing days for wheat cultivation between December and January (https://www.yieldgap.org/accessed on 19 November, 2022). The rainy season begins in October and lasts until April, with most precipitation during December and February (38).

Sampling was performed in July after durum wheat (Triticum turgidum var. L. durum) harvest. Soils were sampled between 5–30 cm. Soil samples were collected from within 2.5 km of Fez, at altitudes ranging from 60 to 300 m. Three different types of wheat field were sampled: i) those planted with Eucalyptus trees as a companion crop (EU); and those that were kept fallow after harvest residue mulching for either ii) long-term (15 months, FAL); or iii) short-term (3 months: FAS). The cropping system of all sampled fields was a rotation of wheat and barley. Three different EU locations were sampled, and two different FAL and FAS locations. Sample locations and names are described in the map of Fez region within the supplemental information. Six subsamples per sampling site were randomly collected from different points of the field, and then pooled. After removing plant and other debris, soil samples were air-dried for two days, sieved with a 2 mm screen, and stored at 4°C prior to laboratory analysis.

Electrical conductivity and pH were measured in a 1:10 (w/v) aqueous solution (39). Moisture content was obtained by weighing before and after drying at 105°C for 12 h. Organic matter (OM) concentration was determined by the loss on ignition method at 430°C for 24 h (20). Total carbon, and N were determined by an automatic C/N/S analyser (NA1500, Carlo Erba, Emmendingen, Germany). Macro and micro-nutrients, and heavy metal content was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES; Thermo Scientific iCAP 6500, Waltham, MA). Water-soluble carbon (WSC) was determined in a 1:10 extract by a TOC analyser for liquid samples (Shimazdu 5050A, Kyoto, Japan) after shaking for 2 h and filtering through ashless filter paper (Albet 145 110). Water-soluble carbohydrates were assayed with a colorimetric method (40) from the same extract with 4 ml of 0.2% anthrone solution in concentrated sulphuric acid. Absorbance was measured in a spectrophotometer (Termo Electron Corporation Hexiosα). Soil humic fractions were quantified by measurement of the 0.1M Na₄P₂O₇-extractable organic contents, via oxidation with K₂Cr₂O₇ and spectrophotometric determination of Cr⁺³ at 590 nm (20).

Microbial biomass carbon was determined with the fumigation–extraction method, with extraction of organic C by K₂SO₄ (41), followed by measurement of extract C content with a TOC analyser (Shimadzu TOC-5050A). β-glucosidase activity was determined by colorimetric estimation of p-nitrophenol (PNP) formed by hydrolysis of p-nitorophenyl-β-d-glucopyranoside (PNG), as described by Eivazi and Tabatabai (42). The same method was used for the determination of alkaline phosphatase activity, with the substitution of PNG by p-nitrophenyl phosphatase as suggested by Tabatabai and Bremner (43). Urease activity was determined by the method of Kandeler and Gerber (44), estimated from ammonia release following incubation with urea in an alkaline buffer.

Phospholipids were extracted from 6 g of soil with a chloroform-methanol solution based on Bligh and Dyer (45), then fractionated and quantified using the procedure described by Frostegard et al. (46) and Bardgett et al. (47). Phospholipids were transformed into fatty acid methyl esters (FAMEs) by alkaline methanolysis, then quantified by gas chromatography (Trace GC Ultra, Thermo Scientific) through a 30 m capillary column (Thermo TR-FAME 30 m x 0.25 mm ID x 0.25 µm film) with helium as the carrier gas. Temperature was initially 150 °C for 0.5 min, then increased by 2 °C min^-1 to 180 °C, and then by 4 °C min^-1 to 240 °C. The fatty acids 10:0, 11:0, 12:0, 13:0, 14:0, i15:0, a15:0, 15:0, i16:0, i17:0, C18:3n3, cy17:0, and cy19:0 were selected to represent bacterial biomass (46–48), and 18:2ω6 was measured to indicate fungal biomass (49, 50). The ratio of bacterial to fungal PLFAs represents the ratio between bacterial and fungal biomass (47). The Gram⁺ specific fatty acids i15:0, a15:0, i16:0, i17:0, and C18:3n3, and Gram^- specific fatty acids 14:1, cy17:0 and cy19:0 were recorded as a measure of the ratio between Gram⁺ and Gram^- bacterial biomass. The ratio of monounsaturated PLFAs to saturated PLFAs is expressed as mono/sat. All results are given in nmol g^-1.

All results of soil chemical properties, soil enzymatic activities, and PLFAs were reported as means from one-way ANOVA, followed by Tukey’s HSD test (HSD, honestly significant difference at the 95% confidence interval). Error bars represent standard deviation. Physico-chemical parameters, carbon fractions, biomass indicators, enzymatic activities, and the relative abundances of all identified FAMEs of sampled soils were subjected to correlation analysis to assess if measured soil properties differed between different management practices. All statistical analyses were conducted in R (51). The authors have defined the p level for each statistical test for scientific clarity.

Soil characteristics are recorded in Table 1 . All sampled soils were basic in pH (8.5- 8.9) with elevated EC. Calcium was the dominant cation element in all samples. Soils from fields under long-duration fallowing management (FAL-1 and FAL-2) contained significantly higher SOC than the other sampled soils ( Table 1 ). Lower total N was observed in soils under short-term fallowing management (FAS-1 and FAS-2) than long-term fallowing management (FAL-1 and FAL-s2) ( Table 1 ). One of fields with Eucalyptus (EU-2) has the lowest total N (0.7 g kg^-1) together with FAS-2. Degraded soil (DS) from abandoned areas were sampled as a baseline, and contained the lowest SOC and total N, while EC and Na⁺ were the highest of all sampled soils ( Table 1 ).

Carbon fractions were significantly influenced by agricultural management ( Figure 1 ). Long-term fallow management (FAL-1 and FAL-2) soils were high in labile carbon parameters (1750–2325 mg kg^-1 for water-soluble carbon and 110 mg kg^-1 for carbohydrate). On the contrary, soils from fields with Eucalyptus (EU-3) had the lowest contents (748.8 mg kg^-1 for water-soluble carbon and 27.9 mg kg^-1 for carbohydrate). Both FAL-1 and FAL-2 contained higher organic carbon content (1.15–1.2%) than the other sites ( Table 1 ), and high amounts of recalcitrant humic C ( Figure 1 ). The FAL-1 soil had the highest humic C content (2.76 mg kg^-1)( Figure 1 ), and total organic carbon content of the different soil samples ( Table 1 ).

Concerning general microbial parameters ( Figure 2 ), the highest value of dehydrogenase (2.2 μg INTF g^-1h^-1) was recorded in FAL-2, followed by EU-2 and FAL-1 (1.76 and 1.45 μg INTF g^-1h^-1, respectively). Microbial biomass carbon was also greatest (237.26 mg kg^-1) in FAL-2, followed by FAL-1 and EU-1 (228.60 and 183.87 mg kg^-1). Regarding extracellular enzymatic activities, Soils of FAL-2 and EU-1 showed high β-glucosidase activity (0.874 and 0.818 μg PNP g^-1h^-1, respectively). The highest alkaline phosphatase activity was recorded in EU-1 (2.53 μg PNP g^-1h^-1). The greatest urease activity was observed in FAL-2 (3.52 μg NH₊₄-N g^-1h^-1) in Figure 3 . The lowest activities of β-glucosidase (0.02 μg PNP g^-1h^-1) and alkaline phosphatase (0.987 μg PNP g^-1h^-1) are seen in FAS-2, while the lowest urease activity (0.26 NH₊₄-N g^-1h^-1) is observed in EU-3. And those two soils have relatively low carbon content ( Table 1 ). The shortage of OC in the degraded soil baseline (DS) was congruent with lowest activities of all three specific enzymes (β-glucosidase: 0.13 μg PNP g^-1h^-1, alkaline phosphatase: 0.10 μg PNP g^-1h^-1; urease: 0.02 NH₄-N g^-1h^-1) as well as markers of general microbial activities (SBC and dehydrogenase).

Different management resulted in significantly different microbial biomass ( Table 2 ). Generally, soils with higher carbon content (SOC > 1 g 100 g^-1) such as FAL-1 and FAL-2, maintained high abundance of bacteria, fungi, and total microbes. These two soils also displayed higher total nitrogen content (1.3 g kg^-1). In contrast, soils with low organic carbon (SOC < 0.75 g 100 g^-1) such as FAS-1, FAS-2 and EU-3 had lower microbial abundances (Bacteria: 15.3–19.6 nmol g soil^-1, Fungi: 0.5–0.8 nmol g soil^-1). Bacteria were more abundant than fungi in all sampled soils. Lower populations of Gram^- bacteria (1.1–1.2 nmol g soil^-1) were observed in soils with low SOC (FAS-2, EU-1 and EU-3). The highest ratio of monounsaturated to saturated fatty acids (0.250) was observed in FAL-2.

The degraded baseline soil (DS) contained the lowest levels of fatty acids (13.4 nmol g soil^-1)), and the lowest abundance of all microbe classifications (Bacteria:13.0 nmol g soil^-1, Fungi:0.4 nmol g soil^-1. Gram⁺: 1.2 nmol g soil^-1, Gram^-: 0.8 nmol g soil^-1, Saturated: 7.3 nmol g soil^-1; Monosaturated: 1.0 nmol g soil^-1). Furthermore, the lowest ratios of monounsaturated: saturated fatty acids (0.137), and Gram⁺/Gram^- ratio (1.5) were observed in the DS soil sample.

A correlation matrix allows summaries of relationships between nutrient content and microbial communities across different land uses. Within sampled soils, SOC was positively correlated (Pearson correlation efficient (r.) =0.819) with total N ( Figure 4 ). General microbial markers (SMB and dehydrogenase activity) were negatively correlated with EC [(r. =-0.483 and -0.492, respectively) ( Figure 4 ), indicating a stressful circumstance for microbial growth. Negative correlation (p <0.05) was observed between EC and different parameters of microbial communities (e.g. fungal biomass (r. =-0.574)), enzymatic activities (e.g. β-glucosidase (r. = -0.646), alkaline phosphatase (r.=-0.727)) and carbon fractions (e.g. humic substances (r.=-0.554). In contrast, SOC was found to be positively correlated with all other variables, except for the ratio of fungi and bacteria. Nitrogen content was also positively correlated with high abundance of different microbial communities and enzymatic activities.

Eighty percent of agricultural land in Morocco is rain-fed (52), and sustainable yet productive cultivation practices are needed to allow reliable food production for an increasing population (8). Increased resiliency against drought, soil degradation (53), and extreme weather events is also required. Previous work with crop simulation modelling (54) has indicated that improved soil moisture will be needed in Morocco to reduce the gap of 5.35 t/ha between potential and realized yields. Different crop rotation management and agricultural practices can be utilized to improve soil fertility, water holding capacity, and crop yield (54). Several studies report implementation of techniques within Morocco such as reduced tillage, soil cover and crop rotation (12, 55), conducive for conservation agriculture. According to El-Shater and Yigezu (12), retention of harvest residue in the field could increase wheat yields by as much as 30% through increased rain infiltration and suppression of weeds.

This study reports how different land management practices in wheat fields affect biologically-available soil nutrients (C and N), microbial communities, and enzymatic activities. Higher SOC (SOC > 1.1%) was observed in the long-term fallow management treatment (FAL-1 and FAL-2), and these soils also conserved abundance of different microbial communities despite semi-arid conditions, due to the availability of different microbial substrates. Total water-soluble carbon (WSC) is considered an index of organic C mineralization and primarily composed of labile and easily degradable compounds (20), while carbohydrates represent energy readily available for microorganisms (56, 57). The labile SOC, such as the WSC and carbohydrate fractions correlated with increased active microbial populations as observed by dehydrogenase activity (58). Carbon and nitrogen content can greatly influence microbial dynamics in semi-arid and arid regions, as the emergence of a new community structure is dependent on soil C to N ratio (19, 59).

Bacterial abundance was higher than that of fungi in all samples, perhaps due to lower fungal tolerance against regional stress including high temperatures and alkaline soils (60). Gram^- bacteria are less resistant against severe environmental conditions (61 and 62) perhaps due to structural differences, including the absence of a peptidoglycan layer. In this study, low abundances of Gram^– bacteria were detected in soils with low SOC (FAS-2 and EU-3). The lowest Gram⁺/Gram^- ratio in the DS field may be due to low nutrient (C, N) availability (63). The low-SOC soils also displayed low enzymatic activities and low soil microbial abundance, similar to the high-salinity DS soil. The lowest Gram+/Gram- ratio in DS ( Table 2 ) is probably due to low nutrient availability (63). The lowest monounsaturated to saturated fatty acid ratio was observed in DS ( Table 2 ), strongly corresponding to limited nutrients and high salinization as a consequence of soil degradation. This ratio may serve as a stress indicator (64, 65), and lower ratios are often recorded in microbial communities inhabiting environments where SOC and/or nutrients are limiting (66). Due to previous field management practices, the DS soil may have become degraded over time resulting in soils unsuitable for crop production (67).

Land use and management can strongly affect soil microbial biomass (9, 68 and 69) and enzymatic activities (27). Concerning hydrolytic activities, β-glucosidase interacts with plant and microbial cellulose decomposition processes. Glucose is a reaction product and can serve as an energy source for microorganisms (28, 70). Urease plays an essential role in the N-cycle, hydrolyzing urea to yield ammonia and CO₂ (29). Phosphatase is important for the mineralization of organic P (30, 71). Phosphatase release (induction) by soil microbes and the capacity to solubilize unavailable phosphate may be agriculturally advantageous, as a global phosphate shortage is predicted within the next 50–100 years (72). These extracellular enzyme activities can be altered by several factors, including land management practices and the physico-chemical parameters of the soil-plant interface within the microbial habitat (57).

In this study, soils managed with long-term fallowing (FAL-1 and FAL-2) were enriched with SOC and displayed high activities of the three enzymes described above ( Figure 3 ). In contrast, soils with low carbon content (such as FAS-2) displayed low activities. As all three activities are correlated positively with SOC and TN ( Figure 4 ), this result may be a response to high or low amounts of substrate within the soil. In comparison, the critical shortage of OC in the degraded soil (DS) resulted in lowest activities of all three enzymes, as well as lowest general microbial activity (SBC and dehydrogenase). There was also negative correlation between markers of general microbial activities (SMB and dehydrogenase activity) and EC ( Figure 4 ). This may indicate a severely limited environment for microbial growth, as high EC and low SOC can trigger osmotic stress (70), and is typically problematic in semi-arid and arid agricultural conditions.

Fallowing contributes not only to the accumulation of SOC,but can also increase levels of other nutrients (73) and water conservation by increasing soil aggregation, which in turn enhances water-holding capacity and reduces water run-off (15, 18, 74). Fallowing of longer durations is preferable for the improvement of soil fertility (18). However, long-term fallowing is not always feasible under certain circumstances (e.g., increasing human population pressure and land scarcity).

In Morocco, a harsh climate and population expansion require crop yields which may exceed the potential of current systems, and necessitate improvement of soil fertility (53). According to previous work comparing different fallowing lengths (17, 75, 76), durations of two years still can result in significant gains. However, optimizing the fallow length for soil properties and crop production should be considered based on various biophysical factors such as soil types, crop types, and climate conditions (77–81) as well as farmer’s livelihood (76).

In the case of the current study, soil treated by a year-long fallowing period (FAL-2) had the highest humic C fraction ( Figure 1 ), in accordance with the highest organic carbon content of the different soils sampled ( Table 1 , statistically equivalent to FAL-1). Fallow vegetation and harvest residue require time to decompose, simultaneously replenishing a portion of the nutrients removed by the crop and protecting the soil against erosion.

Compared to soils managed with fallowing, soils taken from wheat fields mixed with Eucalyptus trees generally contained lower amounts of carbon, enzymatic activities, and microbial communities. The Moroccan Forest Service began to encourage the establishment of Eucalyptus plantations in 1949 - Eucalyptus species now represent 40% of the nation’s total reforested area (14). Mixing farming with trees and crops is common in Morocco (82). Integrating a crop with Eucalyptus trees can modify the microclimate (e.g. reduced air temperature and net radiation distribution), resulting in conditions more favorable to wheat growth by reducing stress during the post-anthesis period (83, 84). This practice can also contribute to soil multi-functionality for primary production, water regulation, climate regulation, soil biodiversity, and nutrient cycling (24). However, local agroecological knowledge about tree-based diversification was scarce among farmers in the study area (82). Furthermore, farmers are mostly not interested in woody species locally classified as “wild trees” including Eucalyptus, as production requires costly permits from the forest authority (82). Additional disadvantages of integrating crops and Eucalyptus include possible reduction of soil moisture and nutrients due to competition (85).

The results of our study indicate that EU management was not associated with elevated soil organic carbon and soil microbial communities. Future studies should additionally assess other metrics including crop yield, use efficiency of resources (e.g. nutrient, light and water), and cost-benefit ratio for calculation of return-on-investment.