Results

Summary values from the regression analysis (Table 20.5) indicate that the model performed well in predicting soil organic carbon sequestration in annuals, banana, and coffee. Over 50% of the variation in soil organic carbon was explained by the model for annuals, banana, and coffee land use showing that the assumption upon which the predictions were based held true. Predictions were more accurate in banana and coffee land use where there are better soil conservation practices (minimizing soil loss) and no tillage (minimizing oxida-tive soil organic carbon losses), unlike in annual crops. The poor performance of the model on range lands may have been due to errors in accounting for the soil organic carbon inputs and losses. It was assumed that all the herbage in the range lands constituted inputs, whereas in reality part of it is grazed by livestock, which led to a net loss of input and/or diffuse redistribution of the soil organic carbon inputs as dung. On

Table 20.5 Predictive Performance of the Model

Land Use

Slope

Intercept

R2

Annuals

-0.30

2.83

0.54

Banana

-0.19

4.29

0.98

Coffee

0.30

2.60

0.70

Range land

-0.07

4.41

0.07

the soil organic carbon losses, erosion was assumed to be the major loss pathway and yet annual sediment yields are low on this land use (Mulebeke, 2002).

In light of the Cs and g values, the model shows that there is a great potential for soil C sequestration in perennial crops followed by range lands compared to annuals.

0 0

Post a comment