Soil Fertility as a Contingent Rather than Inherent Characteristic - Principles of Sustainable Soil Management in Agroecosystems - page 145

controlled by using a simple push weeder that actively aerated the soil as it churned
them back into the soil as a kind of green manure.
One self-evidently beneficial practice was the application to the soil of as much
compost as possible, made from rice straw and any available biomass. While it was
not expected that compost application could promote higher yields than chemical
fertilizer, nevertheless, this result was seen in large-scale, replicated factorial trials
(Uphoff and Randriamiharisoa 2002).* More output was being achieved with fewer
inputs, which seemed illogical.
There was usually a wide range of yield results with SRI management, and some-
times there was no improvement in yield, usually when water control was not main-
tained and the paddy soils were mostly or always anaerobic. With the same methods
used to measure both SRI and conventional paddy yields, the fourfold increase in
average yield could not be attributed to measurement error or to grain moisture con-
tent (Surridge 2004). In this situation, crop yields appeared to have been “decoupled”
from inherent soil fertility as assessed by standard agronomic criteria.
The extreme phosphorus constraint noted above might have been counteracted by
the SRI practice of alternately wetting and drying paddy soils. Such water manage-
ment could mobilize N from atmospheric sources through biological processes and
also available P from the soil's usually large reserves of unavailable P. Alternating
aerobic and anaerobic soil conditions expands and contracts the in-field populations
of soil microbes, including P-solubilizing bacteria (Birch 1958; Magdoff and Bouldin
1970; Turner and Haygarth 2001; Turner 2005). However, research on factors con-
tributing to SRI effects on paddy yield suggested that possibly more was involved
here than microbes enhancing nutrient availability in the soil.
In 2000 and 2001, large-scale factorial trials were conducted in two contrasting
agroecosystems, comparing the effects of SRI management practices, respectively
and collectively, with standard practice. The first set of trials ( N = 288) at Morondava
on the west coast, at sea level, with tropical climate, and on poor (sandy) soils, found
that conventional methods (older seedlings, 3 per hill, with continuous flooding
and NPK fertilization) gave average yields of 2.11 and 2.84 Mg ha -1 from local and
improved varieties ( riz rouge and 2798), respectively. In contrast, when SRI methods
were used (young seedlings, 1 per hill, with intermittent irrigation and compost),
the average yields were 5.96 and 6.83 Mg ha -1 from traditional and new varieties,
respectively (Uphoff and Randriamiharisoa 2002).
* Two evaluations done in the latter 1990s for the French aid agency in Madagascar reported similar
average yields of more than 8 Mg ha -1 when SRI management was used on farmers' fields, where
standard methods gave only 2 to 3 Mg ha -1 (Bilger 1996; Hirsch 2000). The 2000 report showed SRI
methods producing average yields of 8.55 Mg ha -1 from an area that had expanded from 34.5 ha in
1994/1995 to 542.8 ha in 1998/1999. Use of “modern methods with inorganic fertilizer produced an
average yield of 3.77 Mg ha -1 , which was 60% more than the 2.36 Mg ha -1 that farmers obtained during
this same period in the same irrigation schemes with their usual methods. Some farmers who used SRI
methods very rigorously obtained yields in the 10-15 Mg ha -1 range, and a few had even higher yields,
considered unattainable according to prevailing agronomic thinking” (Dobermann 2004; Sheehy et al.
2004). Very high yield with SRI methods has been recently demonstrated in India, with a world-record
paddy rice yield of 22.4 Mg ha -1 measured by technicians and accepted by the Indian Council for
Agricultural Research (Diwakar et al. 2012). The same hybrid varieties cultivated on these same soils
gave only one-third as much yield as did SRI methods.