Assessing soil cracks’ dynamics in relation to mucilage application and soil hydraulic properties under different soil management practices in Wallonia Belgium, using image analysis

Soil cracking is a common phenomenon dominating the agricultural field. Several observers affirmed that this phenomenon can have positive or negative impacts on the soil hydrodynamic including water retention, soil water permeability, and soil desiccation which in turn adversely affect the crop productivity. However, the mechanisms of soil cracking remain unclear and its effect on soil water characteristics is hard to measure and quantify. This is due to the fact that a deep understanding requires both knowledge of crack analysis (image analysis) and soil hydrology. The study is more complex and scarcer if agricultural parameters such as tillage and residue management are added into the research. Yet, knowing soil cracking process under actual agricultural management (tillage + residue management) will help improving crop productivity.
The first major challenge in this research concerns the methodology adopted in order to obtain information on the dynamics of cracking. While crack analysis has been conducted before; detailed crack observations under undisturbed agricultural soils are rare. Digital image analysis (RGB and two-dimensional) is a promising option because it is proven scientifically and technically acceptable thanks to the open-source software ImageJ and PCAS. These software evolve rapidly due to evolution of technology and growing interaction between users and scientific members. However, the quality and the accuracy of the results (cracking) reside mainly on the modality of image acquisition, manipulation and processing. By investigating several techniques specific to crack analysis, we were able to adopt an improved and semi-automated approach that required the use of ROI (Region Of Interest). This consists of converging the image analysis on the probable areas of cracking in order to reduce the contrasting effects of the image background.
The objective of this thesis is to analyse crack dynamics in connection with soil water properties and mucilage application; under contrasting conditions of agricultural practice and residue management. The results showed that the extent of cracking differs significantly among the adopted treatments, namely RTRI (Reduced tillage with residue incorporation), CTRO (Conventional tillage with residue exportation) and DS (Disturbed soil). In addition, disturbing the soil (as right after tillage) increases cracking more than 5-6 times (CIF: 0.3 % for CTRO, 0.5 % for RTRI and 3 % for DS). This shows the importance of the ploughing period as well as the soil structure. Finally, other parameters such as soil aggregation, cohesion of soil particles, organic matter, fibre content, and soil density (porosity) intrinsically affect the cracks formation and propagation. This is the reason why tilled soil (CTRO) exhibited less cracking than reduced tillage soils (RTRI). Cracking curve grows quickly on disturbed samples as they only have two periods (B and C) instead of three for NDS (undisturbed samples). The length and the area of crack are statistically higher on DS >> RTRI > CTRO.
The result also revealed that initiating crack requires stronger negative suction and lower water content in DS>>CTRO>RTRI. Crack progression (in length and area) increases almost linearly with suction until reaching 300 kPa for DS, and more than 15 000 kPa for NDS. The suction above 5000 kPa was extrapolated form fitting curves of van Genuchten (monomodal model) and Durner (bimodal model). The critical water content (around 20% Wc) is attained at the end of the constant rate period (CRP: maximum evaporation). After this period, crack development commences to accelerate on NDS (RTRI and CTRO). The pore size distribution as well as the Krisher’s curve demonstrate the importance of pore > 50µm (and cracks) on soil permeability and likely on evaporation. The retention curve (SWRC) and the pore distribution curve (PSD) revealed the moment of the onset of cracking. Soil structure as well as the agricultural management affect the crack progression. And as cracks affect the fluid movement in the soil, they also should affect the retention and conductivity functions during desiccation. The effect of soil disturbance seems more important than change in agricultural practices as far as crack formation and soil hydrodynamic is concerned.
Finally, some preliminary result on mucilage (Chitosan (CHI), Tragacanth (TRA), and Xanthan (XAN)) application shows that the drying mechanism of soil has been affected by mucilage. Those rhizospheric substances can absorb significantly water and increase the water retention capacity of the soil (up to 75-80 % in volume for TRA and XAN). The use of mucilages (Xan, TRA) at a dose of 3.6g/kg permit a decrease of cracking density as well as the probability entropy by (9 % and 1%), and (59 % and 12 %), respectively. Those two types of mucilage deferred as well the crack initiation by 5h. Xanthan continue to maintain its effect on crack restriction even at lower quantity (0.9 g/kg) while TRA loses its controlling capacity at similar dose. On the other hand, the effect of CHI is almost similar to reference soil (REF) at any dose. By increasing the temperature from 25 °C to 50 °C, crack formation is increasing under REF/CHI while decreasing with TRA/XAN. TRA/XAN reduces the crack disorder and the fractal dimension indicating a decrease of the crack complexity (in pattern).
This dissertation highlights some parameters affecting soil cracking (agricultural practices and use of mucilages), as well as the probable influences of cracking on the hydro-pedological property by means of image processing and analysis.