The cassava should have an effect on runoff and soil loss, however LAI may not be the best way to quantify it. The following article may be helpful; Wentai et al. (2011) The suitability of using leaf area index to quantify soil loss under vegetation cover. Journal of Mountain Science 8: 564-570. Doi: 10.1007/s11629-011-1121-z
A quick answer to your two questions. From the USDA soil loss equation you can see that plant cover is a decisive variable in soil loss (although some soils are hardly erodible even when bare). Moreover, plant cover is the most manageable among the variables in the equation. LAI may be a good proxy for a protective plant cover of crops and lower vegetation, but not necessarily for tree crops and forests. For trees the (drip) height of the canopy is an essential variable not captured (directly) by LAI. However, LAI may give a seasonally wrong impression in case of perennial grass providing protective cover in steppes, prairies and the likes, or of (non-green) plant material covering the ground at the onset of the rainy season.
From a theoretical perspective, cassava may be expected to control erosion depending the crop rotation schedule in relation to the rainy season. Bare ground at the start of the rainy season is not very helpful, of course.
Main characteristics of crop structure that affect soil erosion are a) canopy cover, b) residue and biomass cover above soil and c) water drop height. Most influential parameter in erosión is the surface soil cover, even under a profuse or poor canopy cover. In general terms, Cassava is not an erosion-control crop. Residue level offered by Cassava are not abundant, and, in most cases, soil is tilled as weed control. Soil losses are mitigated when mulch cover or no till are provided. Practices as living edges can be efficient too in controling erosion even on sloped lands.
See: Evaluation of Vetiver edges in a shifting agriculture system on a sloped land (in Spanish).
LAI is not a good index for representing the protection capacity of crops. It does not represent the crop conformation or arquitecture (canopy and residue cover distribution and/or projection over the soil).
However, LAI is a very important parameter for crop growth estimation in erosion simulation models. For example, in models as EPIC, APEX and SWAT, USLE-Factor C, is obtained from above soil biomass and residues, which depend on LAI. In erosion modeling, that is the main linkage between LAI and erosion.
Article EVALUACIÓN DE BARRERAS VIVAS DE VETIVER EN UN SISTEMA AGRÍCO...
Hello Dennis. I do not have an english version of this article. However, I've already uploaded the bilingual presentation that I used in the IV International Conference on Vetiver 2006. Please, download it from the link I gave you, under "SUPPLEMENTARY RESOURCES".
The leaf area influences the erosion, especially in the sheet erosion that is caused by the impact of raindrops on soil. The higher the soil cover, the lower the index of erosion.
Stem-flow is probably another variable that LAI does not capture itself, but would be related. Also, LAI will affect shading of the soil and generation of organic matter, both of which can stimulate soil biota which increases soil macropores and hence, greatly increasing infiltration and reducing runoff. The amount of litter produced by leaves which would be related to LAI, also increases ponding or ‘depression storage’ of water on the soils surface, which also affects infiltration and runoff values.