Currently, agricultural systems are considered (analyzed and studied) from different points of view compared to the past. One is the protection of the environment in terms of carbon emission and soil characteristics ( López-Garrido et al ., 2009 ; Pezzuolo et al ., 2014 , Basso et al ., 2016 ). Soil characteristics are negatively modified by soil compaction, a side effect of modern agriculture experienced on soils in different parts of the world ( Pezzuolo et al ., 2017 ). Soil compaction is defined as “the process by which the soil grains are rearranged to decrease void space and bring them into closer contact with one another, thereby increasing the bulk density” ( Kroulík et al ., 2011 ). Soil compaction leads to negative consequences such as the reduction of soil porosity, decrease of aeration ( McHugh et al ., 2009 ), reduction of saturated hydraulic conductivity and an increase in soil resistance to roots exploration ( Balbuena et al ., 2003 ; Valdes-Abellan et al ., 2015 ). Some different approaches have been recently proposed in the last year for fast characterisation of soil condition, mainly consisting of sensors mounted on tines or discs allowing on the go data collection ( Chukuw & Bowers, 2005 ; Hemmat et al ., 2008 ; 2009 ). A more traditional approach is the measurement of soil mechanical resistance is, assessed taking advantage of a penetrometer with a conical tip. Mechanical resistance, expressed as cone index (CI), is calculated dividing the insertion force by the base area of the cone. Such stop and-go method results a not practical approach in large-scale fields, even when automated, since it is high time consuming and provides only single location variability ( Hall & Raper, 2005 ). Additionally, the method is influenced by soil moisture, which has in general to be considered whenever quantitative analyses are carried out ( Ayers & Bowen, 1987 ; Botta et al ., 2002 ; Hummel et al ., 2004 ). Despite these limitations, the cone index method is relatively simple and intuitive and widely recognised, also by the ASABE (2001). For this reason, such method was implemented for the present study.

Machines traversing fields are the main source responsible for soil compaction ( Chen & Yang, 2015 ), and their most influencing factors are tire dimension, wheel loads, and inflation pressure. Several mechanical, agronomical, and management solutions are available to mitigate soil compaction. Implementation of low ground pressure tires can allow a reduction of soil compaction on topsoil for about one third of the pressure in comparison with conventional practices (50-80 kPa) ( Chamen et al ., 1990 ), and in equipping machines with rubber tracks, a reduction in soil compaction on subsoil is observed ( Ansorge & Godwin, 2007 , 2008 ). It is possible to use lighter machines and reduce passages in the fields adopting minimum tillage or no-tillage techniques. Subsoiling allows to enhance soil porosity and water drainage. In addition, compaction is reduced adding organic carbon on soil, depending on soil texture ( Kumar et al ., 2012 ; Martín-Lammerding et al ., 2013 ).

Generally, soil compaction leads to negative growth conditions for crops due to high mechanical impedance for roots, decrease in soil aeration, and decrease in water storage ( Da Silva & Kay, 1996 ). There are crops more susceptible to soil compaction than others, as suggested by Koch et al . (2008) . Sugar beet ( Beta vulgaris L.) is accounted as a susceptible crop to compaction ( Märländer et al ., 1998 ). Reduced emergence, initial growth, final yields, and root quality parameters are reported in compacted soils ( Chancellor, 1976 ; Gemtos & Lellis, 1997 ; Tolon-Becerra et al ., 2011 ). Compaction can reduce leaf area, dry matter accumulation, and plant population in sugar beet. Furthermore, the total length and distribution of roots in the soil profile can be reduced by topsoil compaction up to 50% ( Brereton et al ., 1986 ). Adopting the previous described solutions, such as using rubber track machines, do not lead to a mitigation of problems due to soil compaction in sugar beet ( Mosimann et al ., 2007 ).

Controlled traffic farming (CTF) is one of the most interesting and often efficient ways to mitigate soil compaction. In CTF, all or most of operations are performed on well -defined traffic lanes. Machines are equipped with satellite guidance systems which permit crossing repeatedly the same lanes; additionally, machine widths are closely matched with standardized track widths >3 m and narrow tires are implemented ( Holpp et al ., 2011 ).

Machines never exit defined traffic lanes, therefore, topsoil is only marginally affected by compaction ( Hamza & Anderson, 2005 ; Chamen, 2006 ). CTF has demonstrated an increase in crop yield related to random traffic farming. Advantages can be significant in root and bulb crop systems for instance potatoes, onions and sugar beet ( Gasso et al ., 2013 ; McPhee et al ., 2015 ). The present work is focused on sugar beet, with the aim of analyzing the subsoil compaction during the growing period of sugar beet with different farming approaches: controlled traffic passages and random traffic.

With regard to penetrometer analyses, Fig. 3 shows average values from different inter-rows after each agricultural operation for the four conditions (CT0, CT3, CT8 and RT).

It can be clearly noted that after the first operation, penetrometer resistances at increasing depths are still very similar, with only negligible differences on the first 5 cm layer. After eight passages (and about after 4 months), the soil has undergone relevant compaction effect, more evident in the case of the CT3 and CT8 condition lanes, but recognizable also in the RT condition. A slight difference can be noticed also in the case of the CT0 condition. This compaction variation is clearly not related to wheels passage but rather to the effect of atmospheric phenomena. The trend can be better appreciated looking at the average penetration resistance at a depth ranging between 5 and 25 cm, where the sugar beet root typically grows ( Fig. 4 ). Specifically, it can be noticed how penetration resistance increases at a rate of about 0.06 MPa after each operation in the case of CT0, 0.10 MPa in the case of random traffic management, and 0.13 MPa in the case of lanes with 3 or 8 passages. In all of the cases, the increase is particular relevant after the first pesticide application. The reason for such behaviour, which is common to all of the scenarios, is ascribable both to the long time passed before the last operation (about two months) but also to the relevant precipitations in the same period.

In order to verify if such soil management difference has a statistically evident effect of sugar beet plant conditions, different parameters were analyzed after harvesting and subjected to ANOVA and Tukey statistical studies. With regard to the regularity of the taproot, no statistical effect was highlighted with an occurrence of forking phenomena which was similar both for the RT and CT conditions.

On the other hand, a significant effect was determined in the case of the taproot circumference. As reported in Fig. 5A , the estimated circumference was higher in the case of un-trafficked lanes (35.4 cm) and lower in the case of trafficked lanes with a minimum in the CT8 condition (27.9 cm). Thus, a compacted soil tends to inhibit the growth of the taproot. Indeed, considering the taproot length ( Fig. 5B ), average values close to 25 cm were detected only in the case of the CT0 condition, while values lower than 20 cm were in general monitored in trafficked and randomly trafficked lanes. The soil compaction ultimately plays a role in both underground and aboveground plants mass. Results relative to above-ground biomass are reported in Fig. 5C with the same trend described above, and the best performance is always displayed by the un-trafficked lanes, with a biomass of about 0.6 kg: 27% higher than in the case of random traffic plants or at CT8 condition. Undisturbed soils and healthy above-ground vegetation allow proper growth of the taproot. This is eventually evident on the taproot mass ( Fig. 5D ) with an average of 1.2 kg, and CT0 sugar beets were 30% heavier than RT ones and 46% heavier than CT3 and CT8.

Soil compaction, in both RT and CT conditions, causes a general decline of sugar beet growth. CT fields showed an improvement only in those rows not undergoing traffic stress. As already stated, such rows are only the 33% of the total area, however, they allow a general improvement of the yield in the whole CT area.

Considering the average sugar beet weights in different scenarios and the relative distribution of rows with different passages (33% as CT0, 53% as CT3 and 13% as CT8), a theoretical total yield increase of 1.3% should be detected in the CT area compared to the RT area. On the other hand, 89.5 and 81.1 t/ha actual yields were found respectively for the CT and RT areas. Such values bring the difference at a difference of about 10%, which is far higher than the theoretical one. Such difference was due not only to an increased production but also to a better condition of the plants in a particularly rainy season. Indeed, CT areas are less subject to losses due to water logging, in a percentage which can vary between 5% and 15% depending on the specific season, weather, and rainfall in particular. Such phenomena affecting more seriously RT areas result in a loss of product which amplify the benefits produced by controlled traffic. Such results are in a good agreement with those reported in literature ( Chamen et al ., 1992 ) relative to European countries, where implementation of CT farming techniques in root and bulb crop systems (such as sugar beet, onions, potatoes, etc.) bring and increased yields of about 4–14%, when compared with RT approaches.

Furthermore, yield difference can be increased by implementing larger machine widths which allow an enhancement of the CT0 incidence on the total area. It is expected that proper application of controlled traffic management can potentially reduce yield losses up to 30%.

The authors are grateful to Dr. Gianluca Rosso for practical support given during the conduction of the experiment.