Fjölrit RALA - 15.06.2004, Qupperneq 135

Solute transfert in an andisol of the French West Indies after application of
KNO3 : from the agreggate to the field experiment
Sansoulet J.1, Cabidoche Y.M.1, Cattan P.2, Clermont Dauphin C.1, Desfontaines L.1,
Malaval C.3
1. INRA, Institut National de Recherche Agronomique, Unité AgroPédoClimatique
2. CIRAD, Centre de coopération Intemationale en Recherche Agronomique pour le Développement,
Département productions Fruitiéres et Horticoles
3. CNEARC, Centre National d’Etudes Agronomiques des Régions Chaudes
Since 30 years, fertilizer supply on banana cropped on andosol in the French West Indies,
takes into account a high leaching of potassium and nitrates, that had been measured on oxisol
and some andosol by Godefroy et al (1975) and Godefroy and Dormoy (1988). The nitrogen
and potassium supply can reach respectively more than 400 kg N ha'1 year1 and 800 kg K ha"1
year"1. Both fertilizers are supplied around the bottom of the banana pseudo-stems. On one
hand, the banana canopy by-passes the incident rainwater into a main streamflow around the
pseudo-stem (stemflow) and a lower dripping flow under leave borders. On the other hand,
andosol allophanes exhibit pH dependent electric charges. The diversity of cropping systems,
including liming, induces a diversity of physico-chemical status. Ranges of observed pH are
frequently under the Zero Point of Charge (ZPC) and these andosol could then develop Anion
Exchange Capacity (AEC) in deep layers which are hkely to retain anions as NO3". With such
characteristics, understanding distributed water flows and consecutive solute leachings under
banana cropping on andosol can help to improve fertilizer supply by minimising solute losses.
We performed 3 types of experimentations :
(i) The AEC and the capacity of the andisol to retain NO3" and K+ have been measured using
the batch method of Wada and Okamura (1980) under different pH conditions.
(ii) Undisturbed soil columns of 1.5 dm3 allowed to study the anion and cation elutions. An
initial amount of solid KNO3 or LiBr has been spread on the top face of the columns, which
has been immediately dissolved during the first V/V0 elution. Discrete elutions were applied
every two days, until 8 V/V0.
(iii) Wick lysimeters, set up at 0.75 m depth, allowed to estimate in situ drainage distribution
and the associated solute leachings, through both high infiltrability layers A (Ksat > 150 mm
h"1) and B (Ksat = 30 +/- 7 mm h"1). The type and length of the wicks have been chosen in
order to collect the same average flow in lysimeters than the flow in undisturbed wet soil; the
HYDRUS 2D model allowed to verify that lysimeters minimized bias. Five replicates of four
positions in respect to the stem positions were set up. The banana cropping received
homegeneous KN03 supplies - 70 kgN(NOs) ha"1, 200 kgK ha 1 - each 3 months.
The breakthrough curves of the B layer for pointed out that 2 phases occured during
elutions: the first one resulted from quick solute elution and the second one from the slow
lixiviation of a sorbed fraction of the NO3" input. K+ showed a similar behaviour. For Li+,
only the first phase occured, without evidence for sorption (fig.l.). The A layer showed usual
behaviour of K+ and Li+ which was characterized by immediate sorption. Nevertheless NO3
exhibited the same two phase behaviour than in the B layer. In both layers, the final sorbed
Br'/Li+ ratio strongly increased while pH decreasing, according to AEC increasing. The batch
experiment confirmed those trends.
111