Comparison between Two Strategies for the Collection of Wheat Residue after Mechanical Harvesting: Performance and Cost Analysis

: The growing population worldwide will create the demand for higher cereal production, in order to meet the food need of both humans and animals in the future. Consequently, the quantity of crop by-products produced by cereal cropping will increase accordingly, providing a good opportunity for fostering the development of the sustainable supply chain of renewable solid fuels and natural feedstock for animal farming. The conventional machineries used in wheat harvesting do not guarantee the possibility to collect the cha ﬀ as additional residue to the straw. The present study investigated the possibility to equip a conventional combine with a speciﬁc device, already available on the market, in order to collect the cha ﬀ either separately (onto a trailer), or together with the straw (baled). The total residual biomass increased by 0.84 t · ha − 1 and 0.80 t · ha − 1 respectively, without negatively a ﬀ ecting the performance of the combine when the cha ﬀ was discharged on the swath. Farmers can beneﬁt economically from the extra biomass collected, although a proper sizing of the machine chain is fundamental to avoid by-product losses and lower revenue.


Framework
The use of non-renewable sources for meeting the fast-growing energy demand worldwide could trigger negative effects on the environment in terms of pollution. On the other hand, as the worldwide population is expected to exceed 9 billion people by 2050 (FAO), the production of several key commodities will also increase accordingly, in order to meet the fast-growing demand for food. The production of cereals is expected to grow from the annual 2.1 billion tons up to 3 billion tons by 2050 if animal feeding is also included [1]. Consequently, the ongoing conflict on land use for food and non-food crops will be more serious if new strategies are not promptly undertaken. Regarding the bioenergy production, the European policy is keen to promote the utilization of agroforestry residues over the plantation of energy crops [2], by applying stringent regulations, in order to meet the climate and energy targets set in the EU 2030 framework [3,4]. Hence, a possible strategy could be improving the collection and the utilization of residual biomass that is normally produced in cereal cropping, but not effectively exploited yet [5]. During the harvesting of cereals, for example, in addition to problems: the removal of weed seeds and the collection of biomass for livestock (both feeding [6,13] and natural bedding). The presence of chaff into straw bales also increases the adsorbent capacity of natural bedding [27]. Even poultry farming can benefit from loose chaff availability on the market. A direct interview with a local farmer in France highlighted the positive effects, noticed by farmers, on the welfare of the animals that could scratch around in search for broken kernels and weed seeds, which, in turn, contributed to overall feeding. The same experience was reported by Italian farmers. The use of similar cereal residues is reported in literature as a valid source for littering. Anisuzzaman and Chowdhury reported that rice husk was a good litter material for rearing broilers [28] and it also has a high adsorbent capacity if compared with sawdust [29]. Chaff could also be suitable for further processing, like briquetting, and used for multiple purposes. Akerlof [30] reported the possibility of producing briquettes of soybean chaff for meeting the needs of livestock in providing complete feeding, whereas spelt chaff has been proven to be a good raw material for the production of briquettes for non-feeding purpose, who exhibited different mechanical properties according to the temperature of compression applied [31]. Wheat chaff applications are not fully studied in the sense of both feeding or not-feeding purposes. The unviability of specific mechanical machines able to collect it without increase in the harvesting costs, has probably limited the research in that direction. For this reason, this study addresses an important issue for the development of new production chains based on cereal residues, showing two possible chaff collection logistics, the limits and operating costs of the technologies used, laying the foundations for the development of possible supply chains that are currently underdeveloped or, in some cases, non-existent. In the framework of the H2020 AGROinLOG project [32], a specific test in the Halland region (Sweden) was carried out, to provide evidences on the possibility of improving the conventional supply chains in wheat harvesting, for increasing the overall residual biomass collectable in the field. Specifically, the aim of the test was to evaluate if it is possible to accomplish such a task by equipping a conventional harvester combine with a dedicated device for chaff recovery, already available on the market and manufactured by the Thierart firm (Thierart, Le Châtelet-sur-Retourne, France) [33]. The device permitted one to flow the chaff, either onto a towed trailer, or on the straw swath produced by the combine harvester. Therefore, both chaff collection methods were tested: loose in a towed trailer (CoT), or baled together with the straw (CoS). The trailer was connected to the combine harvester, therefore no tractor was required for towing it. The amount of biomass potentially collectable as grains, straw and chaff was quantified, as well as the performance and quality of the work of all machines involved in the two supply chains. The loss of seeds, straw and chaff were recorded and an evaluation of the harvesting operating costs was carried out.
Within the field, a homogeneous area of 3 ha was preliminarily selected. The surrounding wheat was harvested and the whole biomass removed, in order to avoid edge effects and biased measurement. The selected area was then divided into three blocks, each of them sub-dived in two rectangular shaped plots measuring approximately 0.5 ha. Thus, three random replications per treatment were obtained, for a total of six plots. The chaff was collected in two different ways (treatments): either discharged on the swath (CoS) or collected on a trailer (CoT).

Pre-Harvest Tests: Theoretical Biomass Assessment
For management reasons, the test was split into two consecutive days: the first day was dedicated to the pre-harvesting activities and combine harvesting; the following day occurred the baling operation and post-harvesting activities. Before harvesting, the whole plants of 10 samples areas of 1 m 2 randomly chosen were hand harvested. Culms and spikes were weighed separately. Successively, all spikes and a representative sample of culms were put in sealed bags and shipped to the laboratory of Research Centre for Engineering and Agro-Food Processing (CREA) for further measurements as: theoretical yield of grain and chaff, dry weight and moisture content.
In the laboratory, by using a stationary thresher (PLOT 2375 Thresher, Cicoria Company, San Gervasio, Italy), kernels were separated from the rest of the spikes (rachis, lemma, glumes and palea). The dry weight and moisture content of culms, kernels and chaff was assessed according to the EN ISO 18134-2:2017 [34] standard.

Equipment
The contractor provided all the machines required for the test. Settings of the combine harvester, as well as the baler, were maintained at a constant rate throughout the experiment.

Combine Harvester and Recovery System
A combine harvester New Holland TX68 with a conventional threshing drum, straw walker and cleaning shoe was used to perform the test. The header was 7.27 m width and it was specifically designed for cereal harvesting. The machine was driven by a 209 kW diesel engine and the chassis was comprehensive of a dedicated hitch for trailer towing.
The device for the chaff recovery was installed at the end of the cleaning shoe of the combine harvester. As shown in Figure 2, the device is made of a tank that receives the chaff from the cleaning shoe; within it, there is a steal-made screw that delivers the chaff to the two-stage turbine which, in turn, blows it through the outlet. According to the company Thiérart [33], the device requires a minimum of 45 L·min −1 of hydraulic oil flow rate to work properly and the cutting bar of the combine harvester should not exceed 5.5 m in width to properly manage the chaff flow.

Pre-Harvest Tests: Theoretical Biomass Assessment
For management reasons, the test was split into two consecutive days: the first day was dedicated to the pre-harvesting activities and combine harvesting; the following day occurred the baling operation and post-harvesting activities. Before harvesting, the whole plants of 10 samples areas of 1 m 2 randomly chosen were hand harvested. Culms and spikes were weighed separately. Successively, all spikes and a representative sample of culms were put in sealed bags and shipped to the laboratory of Research Centre for Engineering and Agro-Food Processing (CREA) for further measurements as: theoretical yield of grain and chaff, dry weight and moisture content.
In the laboratory, by using a stationary thresher (PLOT 2375 Thresher, Cicoria Company, San Gervasio, Italy), kernels were separated from the rest of the spikes (rachis, lemma, glumes and palea). The dry weight and moisture content of culms, kernels and chaff was assessed according to the EN ISO 18134-2:2017 [34] standard.

Equipment
The contractor provided all the machines required for the test. Settings of the combine harvester, as well as the baler, were maintained at a constant rate throughout the experiment.

Combine Harvester and Recovery System
A combine harvester New Holland TX68 with a conventional threshing drum, straw walker and cleaning shoe was used to perform the test. The header was 7.27 m width and it was specifically designed for cereal harvesting. The machine was driven by a 209 kW diesel engine and the chassis was comprehensive of a dedicated hitch for trailer towing.
The device for the chaff recovery was installed at the end of the cleaning shoe of the combine harvester. As shown in Figure 2, the device is made of a tank that receives the chaff from the cleaning shoe; within it, there is a steal-made screw that delivers the chaff to the two-stage turbine which, in turn, blows it through the outlet. According to the company Thiérart [33], the device requires a minimum of 45 L·min −1 of hydraulic oil flow rate to work properly and the cutting bar of the combine harvester should not exceed 5.5 m in width to properly manage the chaff flow. Here, a PVC pipe is connected, in order to permit the discharge of the chaff, either on the swath (Figure 3a) or onto the trailer (Figure 3b). The screw and the twin-stage turbine are driven by the dedicated hydraulic system.
The trailer used was a single-axed wagon, with a pivoted drawbar directly connected to the hitch of the combine (Figure 3a). The loading capacity of the trailer was 6 m 3 . The upper part of the trailer was closed with a thick plastic film, in order to prevent accidental loss of chaff due to wind interference. The combine harvester was also equipped with auxiliary hydraulic connections, for controlling the movements of the trailer while discharging the chaff.

Residual Biomass Harvesting
In treatment CoT, the chaff collected during the harvesting was systematically discharged into an auxiliary trailer parked outside the field, then weighted at the end of every plot, using a local scale. In both treatments, the straw were baled using a round baler John Deer 550 towed by a tractor John Deere 6830. The baler was completely empty at the beginning of each plot. At the end of each experimental unit, the machine was forced to close the bale, even if undersized. The last bale was included in the calculation of the residue production per plot, but not in the calculation of the mean weight of the bales, in order to avoid biased mean weights. In treatment CoS the straw swaths, that also included the chaff, were baled, according to the same methodology applied in CoT. In both Here, a PVC pipe is connected, in order to permit the discharge of the chaff, either on the swath ( Figure 3a) or onto the trailer ( Figure 3b). The screw and the twin-stage turbine are driven by the dedicated hydraulic system.
The trailer used was a single-axed wagon, with a pivoted drawbar directly connected to the hitch of the combine (Figure 3a). The loading capacity of the trailer was 6 m 3 . The upper part of the trailer was closed with a thick plastic film, in order to prevent accidental loss of chaff due to wind interference. The combine harvester was also equipped with auxiliary hydraulic connections, for controlling the movements of the trailer while discharging the chaff. Here, a PVC pipe is connected, in order to permit the discharge of the chaff, either on the swath ( Figure 3a) or onto the trailer ( Figure 3b). The screw and the twin-stage turbine are driven by the dedicated hydraulic system.
The trailer used was a single-axed wagon, with a pivoted drawbar directly connected to the hitch of the combine (Figure 3a). The loading capacity of the trailer was 6 m 3 . The upper part of the trailer was closed with a thick plastic film, in order to prevent accidental loss of chaff due to wind interference. The combine harvester was also equipped with auxiliary hydraulic connections, for controlling the movements of the trailer while discharging the chaff.

Residual Biomass Harvesting
In treatment CoT, the chaff collected during the harvesting was systematically discharged into an auxiliary trailer parked outside the field, then weighted at the end of every plot, using a local scale. In both treatments, the straw were baled using a round baler John Deer 550 towed by a tractor John Deere 6830. The baler was completely empty at the beginning of each plot. At the end of each experimental unit, the machine was forced to close the bale, even if undersized. The last bale was included in the calculation of the residue production per plot, but not in the calculation of the mean weight of the bales, in order to avoid biased mean weights. In treatment CoS the straw swaths, that also included the chaff, were baled, according to the same methodology applied in CoT. In both

Residual Biomass Harvesting
In treatment CoT, the chaff collected during the harvesting was systematically discharged into an auxiliary trailer parked outside the field, then weighted at the end of every plot, using a local scale. In both treatments, the straw were baled using a round baler John Deer 550 towed by a tractor John Deere 6830. The baler was completely empty at the beginning of each plot. At the end of each experimental unit, the machine was forced to close the bale, even if undersized. The last bale was included in the calculation of the residue production per plot, but not in the calculation of the mean weight of the bales, in order to avoid biased mean weights. In treatment CoS the straw swaths, that also included the chaff, were baled, according to the same methodology applied in CoT. In both treatments the fuel consumption registered by the on-board computer of the tractor was recorded for a fuel consumption calculation.

Harvesting and Baling Performance
Every plot guaranteed the formation of four swaths after harvesting, with minimal overlapping between the passes. In treatment CoT, the combine had to stop at least once, in order to empty the trailer and complete the harvesting. At the end of the plot, the trailer was emptied again for total chaff weight. The time required for discharge operations was recorded as accessory time. To measure the grain yield, the collected grain was discharged on a trailer and weighted for each plot.
The performance of the machines was evaluated through the study of the working times, performed according to the Comité International d'Organisation Scientifique du Travail en Agriculture (CIOSTA) methodology and the recommendations from the Italian Society of Agricultural Engineering (A.I.I.A.) 3A R1 [35]. The evaluation of the field speed allowed the determination of the theoretical field capacity (TFC, ha·h −1 ), the effective field capacity (EFC, ha·h −1 ), field efficiency (FE, %) and material capacity (MC, t·h −1 ). Gathered data were used to define the performance of the machines and the operative costs. Fuel consumption during baling was recorded by using the measuring system of the tractor. In the following paragraphs, the biomass unit (t) refers to fresh weight.

Post-Harvesting Test: Biomass Collected, Losses and Bulk Density
After baling, all bales produced within the plots were weighed singularly for total biomass baled assessment and average fresh weight measurement (here, the last bale was not included in the calculation). In treatment CoT, the quantity of the chaff collected was determined by weighing the chaff collected in each plot on an in situ scale.
Losses of biomass were assessed for stubble, straw and chaff. By knowing the cutting height of the combine header, stubbles were reconstructed in the laboratory by cutting the basal part of culms previously harvested for pre-harvest analysis. Straw and chaff losses were determined as the difference between the theoretical biomass available derived from the pre-harvest analysis and the effective biomass weighted at the end of the test. The moisture content of each biomass fraction was measured according to the standard methodology described above. The bulk density (kg·m −3 ) of the loose biomass stored in the trailer was assessed by taking 10 randomly selected samples of chaff and was measured according to ISO 17828:2015 [36]. In each plot, all bales were weighed singularly, and three of them were randomly selected and their sizes measured for volume assessment. Bulk density was successively calculated by dividing the mass in kilograms by the volume in cubic meters.

Cost Analysis
In the economic analysis, the following parameters have been taken into account: purchase and operating costs that were provided by the contractor via a interview, performance of the machines derived from the field tests as primary data, and standard values reported in CRPA methodology [35]. Hourly costs of machines were calculated on the basis of the market value of the agricultural machinery [37,38]. The prices of the machines have been discounted to 2019, applying the lending rate of 3% provided by Banca d' Italia Institute [39]. The parameters used during the cost analysis are reported in Tables 1 and 2. In the calculation of the operating costs of the two harvesting systems, the time required for each operation, the quantity of the products obtained and the respective market value (Table 3) were considered. The economic allocation in each treatment was derived from the ratio between each product revenue on the total revenues obtained, as shown in the following formula: where: Ea = Economic allocation of each product or co-product (i.e., grain seed, straw, or chaff) per harvesting phase (combine harvesting or baling) Mp = Market price of each product or co-product (i.e., grain seed, straw, or chaff) Y i = Yield of each product or co-product (i.e., grain seed, straw, or chaff) R i = Revenue obtained by multiplying Mp × Y i

Statistical Analysis
The statistical analysis was performed in order to discriminate the differences among the treatments. All data were subjected to the analysis of variance (ANOVA), using the R 3.6.1 to separate statistically different means (p ≤ 0.05) [41].

Statistical Analysis
The statistical analysis was performed in order to discriminate the differences among the treatments. All data were subjected to the analysis of variance (ANOVA), using the R 3.6.1 to separate statistically different means (p ≤ 0.05) [41].

Performance of the Combine
The methodologies studied for chaff collection highlighted some differences in the performance of the machines involved. According to what was anticipated by Glasner et al. (2019) [42], the theoretical field capacity (TFC) of the combine did not variate among the treatments, as its speed was constant during the cutting and cleaning processes, although a reduction of 10-25% of cleaning was reported in the study. Despite that, significant differences were found in EFC, FE and MC (Table 4), where the accessory times, like the time required for unloading the wagon in CoT, were included. In

Performance of the Combine
The methodologies studied for chaff collection highlighted some differences in the performance of the machines involved. According to what was anticipated by Glasner et al. (2019) [42], the theoretical field capacity (TFC) of the combine did not variate among the treatments, as its speed was constant during the cutting and cleaning processes, although a reduction of 10-25% of cleaning was reported in the study. Despite that, significant differences were found in EFC, FE and MC (Table 4), where the accessory times, like the time required for unloading the wagon in CoT, were included. In fact, the wagon could collect only 6 m 3 of loose chaff and, considering the low bulk density of 41.75 kg·m −3 , the wagon shortly became full of chaff, forcing the combine harvester to stop and exit the field for unloading the wagon. A similar value of 42.88 kg·m −3 for chaff bulk density was reported by Bergonzoli et al. [26] and slightly higher values of 56 kg·m −3 and 62.08 kg·m −3 were found by McCartney et al. [13] and by Suardi et al. [24]. In CoT, the unproductive times of the combine harvester were 233% higher than in CoS where the chaff was continually blown over the swath. In fact, as depicted in Table 4, the FE and the MC of the combine harvester were significantly higher when the CoS system was applied. In Suardi et al. [19], TFC and EFC were respectively 3.72 ha·h −1 and 2.28 ha·h −1 on average and the combine fuel consumption resulted in 11.8 L·h −1 .
Different methods for loose chaff collection have been reported in the literature. For instance, Suardi et al. [24] tested a continuative discharging of chaff onto a trailer towed by a tractor side by side the combine; that the system permitted to collect 1.27 t·ha −1 of loose chaff. Differently, Bergonzoli et al. [26] tested a combine harvester equipped with Harcob system, which had an integrated tank of 9 m 3 in volume for storage of the chaff collected and that the system allowed to collect 0.6 t·ha −1 . Regardless of the quantity of the chaff collected, neither of them reported negative impacts on the combine performance: the trailer volume available for chaff storage in Suardi et al. [24] was better dimensioned, while the Harcob system allows the simultaneous discharging of grain and chaff, avoiding extra unloading time [26]. For that reason, the unproductive times needed were much lower. Similar tests performed by INRA (Institut National de la Recherche Agronomique) in the frame of the project «Systèmes de Cultures Innovants» and CUMA (Federation Nationale des Cooperatives d'Achat et d' Utilisation de Materiel Agricole) in 2011 and 2012 demodays with similar turbine systems, provided higher results in terms of quantity of chaff collected: respectively, 1.5 t·ha −1 and 1.15 t·ha −1 [43,44].

Performance of the Baler
Regarding the baling stage, the EFC that includes accessories' times (e.g., turning time and unloading time) was lower in CoS, since a higher quantity of biomass in the swath was to be processed (Table 5). That implied more stops for the discharge of the bales and it also forced the tractor to reduce the speed, in order to avoid overloading of the baler's chamber. In fact, the amount of biomass that the baler could process per unit of time was not statistically different. No significant differences were found regarding TFC. The fuel consumption of the tractor was also recorded and referred to the unit of biomass baled. On average, 1.27 (±0.17) l of diesel fuel was required for each ton of straw baled, regardless of the presence or the absence of the chaff in the bales. Table 5. Comparison of the performance of the baler within the two treatments. MC is calculated taking into account the overall quantity of residual biomass produced: straw and chaff together (TFC = Theoretical Field Capacity, EFC = Effective Field Capacity, FE = Field Efficiency, MC = Material Capacity). No statistical differences were found between treatments.

Losses of Biomass during the Baling Stage
The theoretical availability of straw, in the present study, was estimated in 8.02 t·ha −1 ; in line with other studies such as Suardi et al. [24], where the theoretical straw availability was estimated at 7.39 (±0.73) t·ha −1 and 8.33 (±0.75) t·ha −1 , in 2018 and 2019 tests, respectively. Nevertheless, during the present study, the amount of residues baled was on average 3.88 t·ha −1 and 4.68 t·ha −1 with CoT and CoS treatments, respectively (Table 6). Therefore, the remarkable differences in the residue harvesting performance can be imputed exclusively to the suitability of the machine chosen by the contractor, to carry on the baling stage. The round baler John Deere mod. 550 used during the test was equipped with a pick-up 1.41 m wide, whereas the straw swath produced by the combine harvester measured 1.74 m in width, on average. Hence, 0.33 m of straw swath could not be collected by the baler's pick-up system in each pass, due to reduced width of the its pickup system (Figure 5). At the end of the baling phase, a large quantity of product was still not harvested in the field (Figure 3).  When chaff was not included in the bales, the performance of the baler was not statistically different. The fuel consumption ranged between 0.77 (±0.15) l·t −1 and 0.94 (±0.12) l·t −1 in the case of straw and chaff baling, while it ranged from 1.01 (±0.13) l·t −1 and 0.64 (±0.23) l·t −1 , when the chaff was dispersed on the ground. In similar experiment, the TFC and EFC of straw baling operation resulted on average 3.96 ha·h −1 and 2.01 ha·h −1 , with a mean FE of 51 % [19].

Losses of Biomass during the Baling Stage
The theoretical availability of straw, in the present study, was estimated in 8.02 t·ha −1 ; in line with other studies such as Suardi et al. [24], where the theoretical straw availability was estimated at 7.39 (±0.73) t·ha −1 and 8.33 (±0.75) t·ha −1 , in 2018 and 2019 tests, respectively. Nevertheless, during the present study, the amount of residues baled was on average 3.88 t·ha −1 and 4.68 t·ha −1 with CoT and CoS treatments, respectively (Table 6). Therefore, the remarkable differences in the residue harvesting performance can be imputed exclusively to the suitability of the machine chosen by the contractor, to carry on the baling stage. The round baler John Deere mod. 550 used during the test was equipped with a pick-up 1.41 m wide, whereas the straw swath produced by the combine harvester measured 1.74 m in width, on average. Hence, 0.33 m of straw swath could not be collected by the baler's pick-up system in each pass, due to reduced width of the its pickup system ( Figure 5). At the end of the baling phase, a large quantity of product was still not harvested in the field ( Figure  3).
The estimated average loss of residue after baling was 4.75 t·ha −1 (4.68 t·ha −1 for CoS, and 4.72 t·ha −1 for CoT), corresponding to a loss of biomass of 50% on average, without statistical difference between the two treatments.
Bergonzoli et al. [26] reported a similar value when a combine harvester mounting Harcob system (developed for Maize cob harvesting) was modified and used for collecting the chaff in wheat crops, even if the results were ascribed to the cleaning system of the combine harvester.  The estimated average loss of residue after baling was 4.75 t·ha −1 (4.68 t·ha −1 for CoS, and 4.72 t·ha −1 for CoT), corresponding to a loss of biomass of 50% on average, without statistical difference between the two treatments.
Bergonzoli et al. [26] reported a similar value when a combine harvester mounting Harcob system (developed for Maize cob harvesting) was modified and used for collecting the chaff in wheat crops, even if the results were ascribed to the cleaning system of the combine harvester.
Such a level of product losses recorded during the tests exceed the sustainable removal rate of 33% proposed by Unger and Glasner (2019) [23]. For this reason, it could be considered positive from the point of view of soil fertility, even if the economic sustainability is closely linked to the amount of recoverable biomass. Therefore, low collection efficiencies may render the operation of recovering residual biomass economically unviable.
However, the scenarios herein proposed provided differences in both the quantity and quality of residuals biomass collectable from wheat cropping, without affecting the grain yield. Such an aspect is very important; in fact harvesting, along with storage, is the most responsible factor for loss of grains throughout the wheat supply chain [45]. The presence of the chaff included in the bales increased both weight and density of the bales by 7.45% and 7.09% respectively, in comparison with bales free of chaff ( Table 6). Increases of 18.0% in bale bulk density, due to the inclusion of chaff, was reported by Suardi et al. [24], when a similar turbine technology for chaff recovery was used. On the other hand, Suardi et al. reported a non-significant increase in the case of chaff admixing performed with a combi system (manifactured by Rekordverken Sweden AB, Kvänum, Sweden) [19].
The different methods studied, allowed to harvest 4.68 t·ha −1 and 4.72 t·ha −1 of wheat residues by baling chaff and straw together, or by harvesting chaff in the trailer and straw baling, respectively (Table 6), with no statistical differences.

Cost Analysis
In the analysis of the unitary costs, the running cost of each machinery involved in the supply chain is related to the market price [€·t −1 ] of each product and by-product obtained. The performance of the machines contributed to the final calculation of the costs. For instance, the reduction in EFC, FE and MC of the combine harvester (Table 4) found that, when the combine towed the wagon (CoT), it increased the hourly harvesting cost by 3.41%, the cost per hectare by 73.35%, and the cost per ton of biomass processed by 67.73% (Tables 7 and 8), in comparison with CoS. Here, the combine harvester did not waste time to continually stop and unload the wagon. Table 7. Costs for unit of time, surface and ton of biomass processed in CoS harvesting system, considering the productivity and the market price of each product.  On the other hand, when the chaff was blown on the swath, the baler had much more biomass (straw and chaff) to process. In fact, the baler's EFC (Table 5) dropped by 18.23% and the costs per hectare and per ton of biomass processed increased by 21.79% and 101.85%, respectively. The hourly cost for baling did not change (Tables 7 and 8).

Unit
The choice to apply CoS over CoT harvesting method affected both the performance and running cost of the machines. According to Table 9, the harvesting cost per hectare increased by 28.90% (from 152.03 €·ha −1 to 196.05 €·ha −1 ), when the chaff was collected as loose material (CoT). Table 9. Economic calculation cost, revenue and the net gain obtained from the collection of grains, straw and chaff when harvested with the two different methods: CoS and CoT. The same results were obtained by Unger and Glasner in 2019, where the separate chaff collection and supply led to higher costs [23]. However, the overall capacity of CoT system permitted one to collect more biomass per hectare (0.38 t and 0.04 t of grain and chaff respectively), counterbalancing the higher costs. In fact, if considering just the net gain per hectare, CoS permitted to gain only 27.52 €·ha −1 .
Furthermore, in the present study, a market price for chaff of 50 €·t −1 was considered. However, Unger and Glasner [23] highlighted that the potential revenue of chaff could vary depending the final use and market price that can range from 81 €·t −1 to 200 €·t −1 , making chaff separate collection economically viable, and giving the farmer, from year to year, different sales opportunities of the product to more profitable markets.

Conclusions
The cultivation of the cereals is an important source of staple food around the world, and it also produces a relevant quantity of ligno-cellulosic biomass, that can be further exploited in order to improve the economic and environmental sustainability of the whole supply chain. In fact, agricultural residues are gaining more and more interest, due to their considerable availability and their potential content of energy, or as raw material for industrial processes. Cereal straw and chaff collected either separately or baled altogether can be a source of food for animals, particularly in case of shortage, or natural bedding for livestock. In poultry farming, farmers reported positive experiences on the use of loose chaff for littering, since it provides wellness to animals and a good adsorbent capacity. However, possible utilization of chaff is to produce bioenergy. Normally, about two tons of chaff per hectare are available, but still not collected, due to three major problems: unawareness of proper mechanical devices available on the market for its collection, uncertainty on the harvesting system to adopt and the development of a specific supply chain for its exploitation. So far, the literature reports few cases of chaff collection with the specific purpose of weed seeds removal, but it still lacks specific experiments on these machines intentionally used for biomass collection. Therefore, the present study aimed to fill that gap and provide deeper understanding in the possibility to enhance the current wheat harvesting method, in order to improve the quantity of biomass collected by including the chaff. This research analyzed the technical and economic feasibility of two different logistic methods for chaff collection: chaff collected as loose product onto a towed trailer (CoT) and baled altogether with the straw (CoS).
Our results suggest that upgrading a conventional combine harvester with a twin-stage turbine for chaff collection increases the total biomass collected by 0.84 t·ha −1 without affecting the grain yield. Furthermore, the separation of chaff from the straw is performed simultaneously to the cleaning process of the grain and no additional passes of the machine on the field are needed, and further soil compaction is prevented.
Even if our results reveal that the collection of the loose chaff into a towed wagon is more costly than including it into the bales, the market price of the pure chaff should be higher, to offset the extra costs required by the contractor for the collection and handling. Furthermore, it should be noted that the trailer system could be used also for other crop by-products; for instance, collecting finely chopped roughage after a forage harvester, without the strong modification of the combine, reducing the unitary cost of the investment and increasing the quantity of biomass potentially collectable. In fact, the unproductive time in CoT was 233% higher than in CoS with an increase of 43.94 €·ha −1 for the harvesting cost. In addition to the higher costs, loss of revenue may take place in case of inappropriate choice of the machine for accomplishing a specific task. Particularly, the round baler chosen by the contractor could not collect all the straw windrowed by the combine harvester. Although the subject is still under discussion, some authors consider that a residue extraction of no more than 33% is sustainable for the soil fertility. On the other hand, however, an amount of uncollected residue, such as that found during the study (50% of harvesting losses), could negatively affect the economic feasibility of the residue collection phase, questioning the investment in specific equipment. In fact, according to 6results from CoT treatment, where the chaff was not included in the straw, only 3.88 t·ha −1 out of 8.02 t·ha −1 of straw available on the field were baled. Considering the straw market price of 50 €·t −1 , this can be translated into a loss of income of more than 200 €·ha −1 .
Future studies should be focused on the assessment of the sustainability of the chaff collection, in terms of the effect to the soil fertility, carbon dioxide emissions and soil compaction.