Toward a mechanistic understanding of the effect of biochar addition on soil water retention

Biochar (BC) is a carbon-rich product produced by thermal degradation of biomass in an oxygen-free environment, whose application to sediment is said to improve water retention. However, BC produced from different feedstocks and pyrolyzed at different temperatures have distinct properties, which may alter water retention in ways difficult to predict a priori. Our goal is to develop a mechanistic understanding of BC addition on water retention by examining the impact of BC from two feedstocks, poultry litter (PL) and hardwood (HW), on the soil-water retention curves (SWRC) of a uniform sand and a sandy loam (SL). For experiments with sand, BC and sand were sieved to the same particle size (~ 0.547 mm) to minimize effects of BC addition on particle size distribution. Experiments with SL contained the same sieved BC. PL and HW bicohars were added at 2 and 7% (w/w), and water retention was measured from 0 to -4.38 × 10⁶ cm-H₂O.

Both BCs increased porosities for sand and SL, up to 39 and 13% for sand and SL, respectively, with 7% HW BC addition. The primary cause for these increases was the internal porosity of BC particles. While the matric potential for air-entry was unchanged with BC addition, BC amendment increased water retention for sand and SL in the capillary region (0 to -15,000 cm-H₂O) by an average of 26 and 33 % for 7% PL and HW BC in sand, respectively, but only 7 and 14 % for 7% PL and HW BC in SL. The most dramatic influence of BC amendment on water retention occurred in the adsorption region (< -15,000 cm-H₂O), where water retention increased by a factor of 11 and 22 for 7% PL and HW BC in sand, respectively, but by 140 and 190 % for 7% PL and HW BC in SL, respectively. The impact of BC on water retention in these sediments is explained primarily by the additional surface area and internal porosity of PL and HW BC particles.

van Genuchten (VG) models were fitted to the water retention data. For SL where the impact of BC addition on water retention was less significant, a unimodal model fit water retention data well for unamended and BC-amended media: the addition of BC did not create a new class of small pores that could not be described with a unimodal VG model. While a unimodal model fitted the sand data well, a bimodal model was required for all BC-amended sand. The influence of BC type, mass fraction and sediment on water retention models will be discussed.