SUMO: SUgarbeet MOdel for predicting yield and quality

Transcription

1 IRS P.O. Box 32 NL AA Bergen op Zoom, The Netherlands / /vanswaaij@irs.nl SUMO: SUgarbeet MOdel for predicting yield and quality A.C.P.M. van Swaaij, L.M. Withagen, (IRS) T. Schiphouwer (Suiker Unie) and A.B. Smit (WUR)

2 contents Yield and quality forecast present yield forecast 2003 history of yield forecast SUMO: estimations SUMO: factors determining yield and quality SUMO: results conclusions

3 Average yield forecast The Netherlands 2003 Based on SUMO date root yield (t/ha) sugar yield (t/ha) final yield final yield present forecast

4 1 = Zeeuwsch-Vlaanderen 2 = Zeeuwse Eilanden 3 = West-Brabant 4 = Noord- en Zuid- Holland 5 = Oost- en Zuid- Flevoland 6 = Noordoostpolder 7 = Noordelijk klei 8 = Noordelijk zand 9 = Noordelijk dal 10 = Gelderland e.o. 11 = Oost-Brabant 12 = Limburg present forecast

5 Yield forecast individual regions 2003 Based on SUMO Region No. Name 1 = Zeeuwsch-Vlaanderen 2 = Zeeuwse Eilanden 3 = West-Brabant 4 = Noord- en Zuid-Holland 5 = Oost- en Zuid-Flevoland 6 = Noordoostpolder 7 = Noordelijk klei 8 = Noordelijk zand 9 = Noordelijk dal 10 = Gelderland e.o. 11 = Oost-Brabant 12 = Limburg Netherlands Root yield (t/ha) present forecast Sugar yield (t/ha)

6 history untill 1995: periodic harvests 1 sampling site per ha 4-10 samplings/year growth models WUR: scientific model Suikerunie: applied model SUMO synthesis between applied and scientific calibration: yield and weather data Oracle database; start in 1996 national forecast without additional sampling

7 estimations actual weather parameters historical weather SUMO growth stage growth yield forecast quality

8 driving factors: factors adjusting factors: temperature irradiation drilling date growth point date potential growth actual growth final yield moisture* variety, plant number, temperature moisture* harvest, transport, storage and other losses

9 estimations Level 3 - acreage - soil conditions - moisture balance - % drought susceptible - % irrigation Level 2 - sowing date - temperature sum required - growth coefficient - reduction factors variety, plant number - regression coefficient yield Level 1 - quality

10 estimations Quality forecast extractability sugar and molasses average of the Netherlands K+Na = a x DCC + b x Precipitation 1 + c x Variety factor + d α-amino N = e x Precipitation 2 + f x Temperature + g x Variety factor + h

11 Quality and variety estimations Quality improvement of sugar beet varieties, (values adjusted for year effects) 105 Rel. value; 1976= K +Na α-amino-n Year

12 Sowing date factors sowing date root yield (t/ha) sugar yield (t/ha) March 1 March 15 March 31 April 15 April 30 May 15 61,1 9,6 60,5 9,6 58,3 9,2 55,4 8,8 51,4 8,3 45,4 7,4

13 factors Temperature I Temperature sum to date of closed canopy: Polders Holland Northern sand 590 o. days 611 o. days 641 o. days

14 Temperature II factors Effect of average daily temperature during summer correction factor 1,2 1,0 0,8 0,6 0,4 0,2 sugar production root growth 0,0 0,8 0,9 1,0 1,1 1,2 1,3 actual/historic 10-days average temperature

15 Radiation I factors 5.5 Radiation Use Efficiency of root production Efficiency (g/mj) Jun 1-Jul 1-Aug 1-Sep 1-Oct 1-Nov Date

16 Radiation II factors Radiation Use Efficiency of sugar production 1.0 Efficiency (g/mj) Jun 1-Jul 1-Aug 1-Sep 1-Oct 1-Nov Date

17 Radiation III factors Efficiency (g/mj) Radiation Use Efficiency of sugar production, difference between areas Flevoland W.Brabant O.Brabant Date

18 factors Variety use of varieties with high sugar content use of resistant varieties with lower yield potential

19 factors Moisture temperature evaporation precipitation radiation moisture balance reduction factor ( ) (per day)

20 Harvest losses factors Actual delivery at the factory (tons/ha) Estimated growth and final harvest y = 0.74x R 2 = Estimated growth until October 12 th (tons/ha)

21 Not included in SUMO factors pests and diseases soil structure fertilization resistant varieties extreme weather conditions These are partly covered by overall regressions over the years.

22 Mid-August sugar yield predicitons (tons/ha) results Year prediction realized difference prediction - realised SUMO sampling yield SUMO sampling average average

23 conclusion forecast SUMO or sampling same precision SUMO: input required is minimal extreme conditions or diseases are not included