Soil respiration plays a major role in global carbon (C) cycle (Schlesinger and Andrews, 2000; Gougoulias et al., 2014) and is actively related to soil ecosystem productivity. Studies identify conservation or sustainable agriculture as an effective option for enhancing C storage, and in some cases, minimizing CO₂ loss from agricultural soil (Martínez-Mena et al., 2021; Jia et al., 2021). Typically, conservation or sustainable agriculture is the culmination of strategies for cultivating crops with minimal inputs, soil disturbances, and implementation of continuous soil cover (mulch) (Hobbs et al., 2008; Zhang and Peng, 2021). Cover cropping is recognized as one of the sustainable agricultural practices that influences soil C balance depending on soil types, land management, and environmental conditions. Ecosystem services provided from cover cropping include organic matter addition, soil structure improvement, weed control, nutrient management, and pollinator attraction, which are especially important in organic, low input systems (Scholberg et al., 2010). Sunn hemp (Crotalaria juncea L.) (SH) and velvet bean (Mucuna pruriens L. DC.) (VB) are leguminous cover crops (CC) with potential to contribute 4 to 14 Mg ha⁻¹ soil organic C (SOC) in tropical climates (Wang et al., 2012; Rosolem et al., 2016).

When CCs are incorporated into a no-till farming system, they have the ability to enhance soil biodiversity and fertility through increasing soil CN fractions and improving aggregate stability (Rigon et al., 2020). While the addition of CCs can be beneficial for soil fertility and crop production, enhanced microbial activity in CC systems can also increase CO₂ emission from soil (Sanz-Cobena et al., 2014). In fact, CO₂ emissions can even be heightened with the use of legume CCs due to the low C:N ratio of plant tissue produced (Alluvione et al., 2010; Muhammad et al., 2019). Therefore, a soil C balance is needed to elucidate both positive and negative aspects of these CCs in a fruit production setting. This research was conducted in an organically produced carambola (commonly known as starfruit (Averrhoa carambola)) orchard, which are commercially grown only in Florida within the mainland United States.

Traditionally, poultry manure (PM) is commonly applied as fertilizer in organic tropical fruit production systems (Canali et al., 2004; Jones et al., 2020). As over 10 billion kg of poultry litter is produced annually in the United States, composting poultry litter is a useful way to repurpose problematic waste into a nutrient source for crop production (Reddy et al., 2007). Repurposing the waste for agricultural purposes is a sustainable method of adding carbon (C) and nitrogen (N) to soils (Nyakatawa et al., 2001; Reddy et al., 2004), while reducing the risk of N leachate polluting natural systems (Kaiser et al., 2009). Because carambola requires intensive nutrient management during production (both at young and mature stages; Crane, 2016), growing of CCs combined with PM additions could be beneficial to soil health and quality, along with providing economic benefits to the farmers, making this a potentially ideal management strategy for tropical fruit producers.

We selected SH and VB as our study CCs because of their large biomass producing nature (average 14–21 Mg ha⁻¹). These CCs have also proven to accumulate large amounts of SOC and to enhance soil health in South Florida (Wang et al., 2015). However, a comparative analysis of C budget from these 2 CCs has not been studied, specifically, in organic cultivation where SOM is expected to play a major role in the C balance. We also recognized that no literature is available on analyzing the effect of CCs on soil respiration under organic no-till carambola production in the US. Carambola contributes $3.7 million per year to Florida’s economy with much room for growth (Ballen et al., 2020) with an increasing popularity for the growers since organic avocado (major fruit crop in South Florida) production has been facing serious pest and pathogen problems throughout the last 10–12 years (Menocal et al., 2018). Adding CCs can apply to many tropical fruit grove settings and results from this research may lay the foundation for this practice to be utilized throughout South Florida. Therefore, the specific objectives of this study were to: a) evaluate the emissions of CO₂ from SH and VB plots (with and without manure treatments) and b) develop C budgets for both CC plots and elucidate the ability of SH and VB to mitigate CO₂ emissions.

Emission of CO₂ from experimental plots were separated for two timescales: A) CO₂ emission from the day after crop germination till the day of CC termination (90 days) and B) CO₂ emission from the day after CC termination (or the day manure was applied) to the end of this study (167 days). Average CO₂ emission (µmol m⁻² s⁻¹) from the control plot (8.22) was similar to SH (8.85) throughout the CC growing seasons (Figure 2). However, CO₂ emission from VB plots (10.21) was 23 and 15% higher than control and SH plots, respectively. The amount of CO₂-C emission during the last few weeks (55–90 DAG) before CC termination was about 42% higher from both SH and VB plots than initial weeks of CC establishment. This is specifically because higher root respiration during full crop growth stages increased CO₂ emission from the soil. Now while comparing the treatments after CC termination, we observed that VB with manure (VBM) plots produced the highest amount of CO₂-C followed by SHM (Figure 3).

Carbon dioxide emission was observed immediately after manure application and high CO₂-C emission peaks were observed (as high as 16 μmol m⁻² s⁻¹) within first couple of weeks after CC termination (or the manure was applied). In this study, 41% of the total CO₂ emission was found within 20 days after crop termination (DAT). Overall, CO₂-C emissions from manure applied plots (average 8,030 kg ha⁻¹ season⁻¹) were 17% higher than plots without manure treatment (average 6,870 kg ha⁻¹ season⁻¹) and the emission was significantly higher (p < 0.05) during 20 days after manure application.

Furthermore, we need to consider both positive and negative aspects of CCs in CO₂ emission and total C contribution (from crop biomass analysis) during production. Positive C balances were found for both SH and VB (Table 2). An average 13200 kg ha⁻¹ TC was contributed by SH, 33% higher TC than VB and comparing all plots, about 8–10% higher (not significant at p < 0.05) TC contribution was observed from manure applied plots than plots where manure was not applied. The average fraction of emission vs C contribution (CO₂-C/TC) was 1.46 times higher in VB plots than SH plots. Higher C emission fraction from VB plots was also an indirect indication of lower C use efficiency by VB cropping system as compared to SH.

Average annual rainfall was 13.25 cm in season one and 11.13 in season two, with relatively higher rainfall observed during July- August (average of 18–24 cm per month) in both years as South Florida experiences a wet summer season and a dry winter season. Rainfall and water filled pore space (WFPS, %; range: 17 to 74, average: 36) did not have any effect on CO₂ emission from all plots, however, soil and air temperature had a significant effect (p < 0.10). For instance, higher CO₂ emission peaks (at 13, 26, 39, 54 days after germination; DAG) were correlated to higher air and soil temperatures.

Higher soil and air temperature have been reported to increase root and microbial respiration (Huang et al., 2005). Similarly, increased CO₂ emissions were observed with higher soil and air temperatures in previous studies conducted in the sub-tropical US (Wang et al., 2013; Dattamudi et al., 2019; Ray et al., 2020). Daily average CO₂-C emission (94 kg ha⁻¹) during CC growing season from our experiment (Figure 2) corroborate well with other CC studies amended with organic manure (Ray et al., 2020). However, our study noted increased CO₂ emission (3–4 times higher) when compared to other studies conducted in mineral soils in the same agroclimatic region (Tian et al., 2015; Dattamudi et al., 2019). This comparative analysis suggests that OM played a significant role (our soil had 17% OM) in increasing soil microbial activity and root respiration during crop production. This can be compared to a recent study in which Bosco et al. (2019) reported that cumulative and daily CO₂ emissions from organic systems were higher than an integrated cropping system for vegetable production. Additionally, lower CO₂-C emissions were observed from fallow (control) plots than CC treatments. This result was likely caused by lower root respiration from those plots. Crop biomass production and ground cover of SH was higher than VB, which resulted in lower soil temperature during SH production, and subsequently, a lower CO₂ emission from soil during the CC growing season. Moreover, the quality and quantity of harvested CC biomass can also have an effect on soil respiration rates (Nilahyane et al., 2020), a possible factor impacting our results.

Furthermore, prior findings have supported that application of chemical fertilizers have little to no effect on CO₂ emission from mineral soils (Dattamudi et al., 2019; Sistani et al., 2019). In our study, CO₂-C emissions from manure applied plots was higher than those that were not treated with PM, specifically within the first 20 days of PM application. The spike in CO₂ emission from manure applied plots possibly resulted from higher microbial degradation of 17% SOM along with increased root respiration in our soil. Average C:N ratio (mol:mol) of SH and VB leaf biomass were in the range of 12:1 to 17:1, similar to previous studies (Valenzuela and Smith, 2002). A C:N ratio lower than 15:1 indicates a net N mineralization in the soil (Hagemann et al., 2016) which likely resulted in higher microbial activity in manure and CC treated plots. Therefore, CO₂ emission peaks were observed immediately after manure was applied (increased microbial activity and root respiration) in both crops. Large amounts of nutrients from CC biomass are often released 30–60 days after incorporation into the soil, which is highly dependent on the C:N ratio of the CC species (Lynch et al., 2016). As a result, a few CO₂ emission peaks observed during 40–75 DAT were likely caused by an increased nutrient release in the second month after CC termination. Our findings indicate that soil respiration (and CO₂ emission) can be influenced by the addition of macro and micronutrients if abundant food sources (SOM) are present for the microbial community. Interestingly, unlike the CC growing seasons, soil and air temperatures had very little effects on CO₂ emissions after CC termination except for few relatively higher peaks (74 and 141 DAT) that corresponded to higher temperatures.

Higher TC contribution by SH (with and without manure) indicates that SH could be a stronger CC candidate for contributing C in organic rich soil than VB. In a study conducted in similar agroclimatic zone, Wang et al. (2012) found lower TC contribution by SH (8.71 Mg ha⁻¹) and VB (4.73 Mg ha⁻¹) monocultures in mineral soil, confirming that organic matter plays a major role in biomass production and subsequently TC contribution (our soil has 17% SOM) in soil. Soils in manure applied plots had higher TC regardless of the type of cover crop. It should be noted that average C content of the PM was 33.76 ± 0.29%, which would amount to 561 ± 67 kg ha⁻¹ of TC contribution, which could account for higher TC contribution in manure applied plots. Interestingly, total C contribution by SH was greater than VB, but CO₂ emission from VB was higher than SH plots. This indicates that C contributed by VB was likely in labile form which stimulated higher microbial activities and subsequently greater CO₂ emissions. Another possible explanation could be the rate of crop biomass decomposition from SH and VB plots. Higher crop residue decomposition was reported to increase soil CO₂ emission (Campos et al., 2011). We did not measure the biomass decomposition rate for SH and VB, but in a recent study conducted in tropical area of Brazil, Xavier et al. (2017) categorized SH as a cover crop with great resistance to decomposition, while VB was categorized as intermediate resistance.

We understand that the soil C balance developed in this study was only from two crop growing seasons, and a long-term field experiment would be more suitable to determine the CC contribution of soil C dynamics. However, in comparing the short-duration effectiveness of SH and VB in contributing/accumulating TC in the soil and influencing CO₂ emission, our result suggest that SH has greater potential to reduce CO₂ emissions from the soil. When sufficient C present in the soil, the addition of N and other nutrients combined with meteorological parameters can significantly influence soil respiration.