1. CGMS Anhui & Yield estimation with RS CGMS Anhui & Yield estimation with RS

2. Our work We participate the following work package  WP 21 Ground data collection  WP 24 CGMS pilot in CHINA  WP 41 Official yield data collection  WP 44 Wheat Yield estimation based on remote sensing for HUAIBEI Plain  WP 7 Networking and Sustainable partnership

3. Our work Organizing a CGMS workshop in China, 2011

4. Part 1 CGMS-Anhui

5. Level 1 Weather Station

6. Level 1 Weather Station

7. Level 1 Update the METDATA The data from meteorological department (Archive data, from 1990 to 2012) The data from NOAA GSOD, now we can download the real-time data from the NOAA GSOD FTP everyday.

8. Level 1 Interpolation Grid Weather The batch model give us a easy way to interpolation weather

9. Level 1 Grid Weather (Average daily temperature, 31/12/2012 )

10. Level 2 Crop simulation- using the batch model 1.Calculate the crop yield 2. Aggregation the grid yield

11. Level 3 Yield Forecast 1.Aggregation the Nuts yield 2. Prepare for forecast

12. Level 3 CGMS Statistical Tool

13. Level 3 CGMS Statistical Tool But only use the potential yield storage to estimate the crop yield, the result is not very good

14. Resent works and further works Update the CGMS dataset Prepared some NDVI, fAPAR and DMP data, plan to add these data into CST Integrate the CGMS Anhui (further work)

15. Part2 Prediction of Wheat Yields Using Multiple Linear Regression Models in the Huaibei Plain of China Beier Zhang (AIER - China ) Qinhan Dong (VITO - Belgium)

16. Contents  Study area  Phenology  Trends of yields  Data sets and methods  Results of prediction  Validation  Discussions

17. Study area Huaibei Plain (include 6 prefectures) Area:64154 km 2 Arable area: 20905 km 2 Main soil type :Cambosols & Vertisols Main crop type: Winter wheat & Maize

18. Phenology SowingEmergenceTiller Wintering period Turning green JointingHeadingMaturityHarvest 10/1210/1912/112/202/103/104/225/156/1 Wheat: October to next year June Maize or soybeans: June to October

19. Trends of yields the trend must be considered There are significant yearly trend of wheat yield in every prefectures from 2000 to 2011, so the trend must be considered in the prediction

20. Data sets I. Biophysical variables based on RS: using SPOT-VGT  Ten-daily series : every dekad from 1999 to 2012  Variables: Smoothed k-NDVI  Building data sets of RS:  The cumulative NDVI for all possible combinations (at least 2, at most 9, because the one phenological stage is less than 3 month) of consecutive dekads within the wheat growing period (2 nd dekad of Oct to 3 rd dekad of Jun).

21. Data sets For example YearO2O3……O2O3O3N1……O2N1O3N2…… 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

22. Data sets III. Meteorology data sets  Variables: include rainfall, temperature and duration of sunshine, from 1999 to 2012  Interpolation method: CGMS Level-1 give us the values of every grid (25km x 25km) in the study area.  Calculate average values in every prefecture  Building data sets of Meteorology data sets:  The average rainfall, temperature and solar radiation of every phonelogical stage of wheat in every prefecture.

23. Methods  Detrend method.  We use two different methods 1.Add year as a variables into the model. 2.Separate the trend yield from real yield, and build the regression model with ΣNDVI and residual error  First predicting the yield using regression to obtain the inter-annual trend (P T )  Calculate the residual error (official yield - P T )  Using ΣNDVI and meteorology data to predicting residual error(P R )  P T +P R

24. Methods  Precision validation  Leave-one-out (leave one year data out; regression model building using the rest of data to predict the left year; corellating the official yield with the predicted ones)

25. Results Prefecture Models R2R2 Absolute Error ConstantyearΣNDVIMeteorology Bengbu+0.645+0.101*year+2.099*O2N2+1.152*F2F30.630 0.372 Bozhou+3.414+0.244*year+2.417*N10.823 0.224 Fuyang-0.156+0.236*year+4.606*O3N1+0.257*Sg0.822 0.278 Huaibei+0.101+0.174*year+0.621*J3M3-0.284*Sj0.831 0.363 Huainan+0.854+0.305*year+0.287*Sg+0.296*Em0.838 0.322 Suzhou+2.118+0.151*year+0.527*J2M3+0.064*Es0.778 0.277 Regression models Using year ， k-NDVI ， and Meteorology Data

26. Results-detrend PrefectureTrend ModelR2R2 Bengbu3.580+0.214*year0.667 Bozhou3.97+0.275*year0.912 Fuyang3.765+0.218*year0.741 Huaibei3.985+0.239*year0.838 Huainan2.859+0.283*year0.751 Suzhou3.900+0.196*year0.754 Trend

27. Results-detrend Prefecture Models R2R2 ConstantΣNDVIMeteorology Bengbu-7.1781.724*M2 +0.721*Sj+0.201*Em+0.293*Ee+ 0.129*Sm 0.667 Bozhou-2.304 +0.279*Sm+0.101*Em+0.298*Ew 0.753 Fuyang-3.720+4.575*O3N1+0.244*Sg0.681 Huaibei-1.147+0.01*J3M3+0.263*Sm-0.152*Se0.585 Huainan-2.877+3.047*N1+0.328*Sg+0.283*Em0.736 Suzhou-257+3.135*O2+1.414*M3+0.331*Ee0.592 Regression models after detrend

28. Validation Using Jack-knife method, comparing absolute error of different methods Prefecture Year & k-NDVI &Meteorology k-NDVI & Meteorology after Detrend Bengbu 0.3720.167 Bozhou 0.2240.159 Fuyang 0.2780.238 Huaibei 0.3630.289 Huainan 0.3220.276 Suzhou 0.2770.27 Mean Error 0.3060.233

29. Validation After detrend Bengbu

30. Validation After detrend Bozhou

31. Validation After detrend Fuyang

32. Validation After detrend Huaibei

33. Validation After detrend Huainan

34. Validation After detrend Suzhou

35. Validation An example,If we want to estimate the yield of 2012. Building the trend model using the data from 2000 to 2011 Calculating residual error. Building the model using the above variables. Then Calculating the yield.

36. Validation The result of year 2012. PrefectureReal yield Estimate yieldabsolute error Bengbu 5.946.210.27 Bozhou 7.537.640.11 Fuyang 6.486.760.28 Huaibei 7.177.580.41 Huainan 6.025.930.33 Suzhou 6.316.790.47 Mean Error 0.31

37. Discussions The method We think the method using k-NDVI & Meteorology after detrend is better This method consider the fact of yield trend, RS and Meteorology. The average error of six prefecture in Huaibei Plain is about 0.233 ton per ha, this is a quite good result.

38. Discussions Suggestion for further study Add these data into CST Add a new dataset from CGMS Level2 Do some field work, get the real crop yield about the field level, then build the model of this level. This work I think can adjust our method and make the result more accurately.

39. Thanks Grazie Merci Bedankt شكرا 谢谢