Capturing near-real time weather data linked with automated biophysical models provides an incremental stream of analysis to the decision-maker, allowing the assessment of emerging trends. In the case of the LEWS program, information is provided in a 10-day interval. This information includes deviation from long-term average of standing crops forage availability to cattle, sheep, goats, camels, and donkeys, the percentile ranking of that response, and the estimated amount of forage available for each target herbivore. Examples of emerging trends in the state of livestock forage that can be seen in the LEWS products are shown in figures 22.3 and 22.4.
Using strong shifts in percentile ranking of current standing crops relative to 30-year averages from generated weather data and changes in NDVI satellite greenness data, it was possible to project forward 30-120 days with a relatively high confidence, allowing decision-makers time to begin planning for adjustments in livestock numbers or movements.
The LEWS program has recently developed a collaborative relationship with the Drought Monitoring Center in Nairobi, Kenya to integrate the quarterly Climate Outlook Forum (COF) 90-day projections of above, below, and average rainfall conditions as well as the 10-day and 30-day projections. Each 10-day report projects 90 days forward using the 25, 50, and
January January January January January January
Figure 22.3 An example of one of the Livestock Early Warning System's products that reflects the percent deviation in forage standing crop for cattle compared to 30-year means of standing crop for a monitoring point in southeastern Kenya.
75% ranking of the corresponding 90-day interval from that date based on 90-day accumulated rainfall from the 30-year generated weather data.
A new promising projection technique developed by Al-Hamad (2002) has been applied to LEWS, where point-based biophysical simulation of forage production coupled with 8-km Advanced Very High Resolution Radiometer-NDVI data was used as a forecasting method for near-term forage production. NDVI data is de-noised using wavelet method (Percival and Walden, 2000) and is then co-regressed with standing crop of forage at the monitoring sites using the Box and Jenkins model or ARIMA procedure for forecasting. This methodology appears to offer stable projections of forage conditions typically from 60 to 90 days.
Was this article helpful?