Agricultural Forecasting

Agricultural Forecasting: Land Use Efficiency: (http://www.esa.int/esaEO/SEM0IM3VQUD_economy_0.html) Informing management decisions Increasing the accuracy of agricultural forecasting is an important application of Earth Observation, informing decisions on agricultural management including irrigation, planting, irrigation, pricing and regional need for food assistance if a harvest is likely to fail. Besides returning information on the impact of prevailing weather, satellite data can also provide a timely summary of acreage under cultivation, potentially differentiating between different crop species. They can also determine plant health, biomass density and soil moisture.

Operational Forecasting of South African Sugarcane production: (http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T3W-4JHMJTT2&_user=10&_coverDate=01/31/2007&_alid=1072861051&_rdoc=2&_fmt=high &_orig=search&_cdi=4957&_sort=r&_docanchor=&view=c&_ct=10528&_acct=C 000050221&_version=1&_urlVersion=0&_userid=10&md5=02892630aecc1a6e5 acd33c6e4f55ccd)

Factors affecting crop yield: • • • • • •

Distance from equator Height above sea level Rainfall Irrigation Growing season per crop (can vary between 12 months and 24) Climatic variability

Between 1980 and 2003 the coefficient of variation for production was 17%.

Model Description: The Canesim model is a daily time step, point-based simulation model predominantly driven by water. For input, it requires soil available water holding capacity (TAM in mm) and daily temperature, rainfall and reference evaporative demand as calculated by McGlinchey and Inman-Bamber (1996). The model

accounts for partial canopy conditions and soil water content using a single layer soil profile. Yield is calculated as a function of transpiration. The water balance of Canesim is described by Singels et al. (1998), canopy development is described by Singels and Donaldson (2000) and the yield calculation by Singels et al. (1999). These publications also report on the validation of various aspects of the model against observed data. The model was further validated at a mill level by Gers et al. (2001), who reported excellent agreement between simulated and observed yields. These results indicate that the science of the model is sound and that it is sufficiently accurate to be used for the purpose of this study. A table list the values for the key factors for each of the areas considered in the study. These included: • • • • • • • •

Total area (km2) Mean annual heat units (oC d an-1) Mean annual precipitation (mm an-1) Mean annual solar radiation (MJ m-2 an-1) Mean coefficient of variance (%) Mean relative discrimination index (%) TAM- mean available soil water holding capacity (%) Mean cane age at harvest

Irrigation Data: For irrigated crops the TAM is allowed to drop to half of that compared to nonirrigated crops.

Weather Data: Weather data was recorded from 32 meteorological stations and 62 rain gauges to provide the most accurate readings for each HCZ Climate stations generally record: • • • • •

Daily rainfall Solar radiation (or sunshine hours) Relative humidity at 0800 and 1400 (dry and wet bulb temperatures) Wind run Minimum and maximum temperatures

Using these sets of data, more useful adaptations can be achieved using Spitters et al., 1986 and Allen et al., 1998 Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration – guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome, Italy, 301p.Allen et al., 1998. As well as the Penman–Monteith equation according to McGlinchey and Inman-Bamber (1996). Data sets were sometimes missing due to errors at the meteorological stations or rain gauges.

System Outputs: The result for the yield of the crop was given as a percentage of the previous year’s yield.

System Evaluation: The accuracy of the system was determined by comparing the predicted yields with the actual yields for 22 years of data. The model consistently over estimated the yields.