Hi Bahram,

I quite agree with Jeffrey’s suggestions.

In most cases, regression analysis is a good method to validate the model’s outputs. In addition, you may consider using other “difference statistics”, such as RMSE, EF etc.
Read following two references for details.

Yang, J., Greenwood, D.J., Rowell, D.L., Wadsworth, G.A. & Burns, I.G. 2000. Statistical methods for evaluating a crop nitrogen simulation model, N_ABLE. Agricultural Systems 64 (1), 37 – 53

Yang, J.Y., Huffman, E. C. 2004. EasyGrapher: software for graphical and statistical validation of DSSAT outputs, Computers and Electronics in Agriculture 45, 125 ‑132.


Jingyi

Jingyi (Bill) Yang, Ph.D.
Soil Science and Modelling
Agriculture and Agri-Food Canada/Agriculture et Agroalimentaire Canada
Telephone/Téléphone: 613-759-1412
Facsimile/Télécopieur: 613-759-1924
Room 4105, K.W. Neatby Building, 960 Carling Avenue
Ottawa, Ontario K1A 0C6
[log in to unmask]



-----Original Message-----
From: DSSAT - Crop Models and Applications [mailto:[log in to unmask]] On Behalf Of Bahram Andarzian
Sent: February 28, 2006 12:11 AM
To: [log in to unmask]
Subject: Variance Analysis

   Dear all my friends
  Hi
  When a model run for different treatments, such as amounts of nitrogen, the simulated result there are in one replication (output). If we going to do variance analysis and comparison between treatments based on traditional statistical methods, it is impossible (no replications). So, is there statistical method, which can be used for evaluating differences among treatments based on simulated data?
  Best Regards
  Bahram Andarzian


---------------------------------
Yahoo! Mail
Bring photos to life! New PhotoMail  makes sharing a breeze.