Statistical down-scaling, because:

1. Cheaper computational resources

2. Provides point-resolution climate variables

3. Direct inclusion of observations - possible

Thanks,

So, what are benefits of dynamical downscaling ?

What does means that SD is cheaper ?

Can SD use to predict ENSO (La Nina and El Nino) ?

Regards,

1. Cheaper in terms of computational resources required.

2. ENSO - may be DD

These steps I should do to get projections of future climate:

1) Download global future climate...

(Where I can get global future climate data?)

2) Downscaling data either dynamical or statistical

Is that right ?

If your rice growth model is very sensitive to meteo uncertainties, you should choose some statistical DS. Dynamical DS is better, if you want to have the physical effects of changing climate of the chosen GCM in your study, but you may need corrections, which adds another layer of uncertainty. For hydrological applications, Wilby et al. (Geophys. Res. Lett. 27, 1199, 2000) or Haddeland et al. (HESS 16, 305, 2012) and others have summarized the effects of downscaling methods. You can also find some hints in my publication in HESS (see my page).

Downscaling techniques are relatively uncomplicated and are adopted to downscale the climate parameter like rain fall or temperature etc These are taken corresponding to a coarse GCM grid cell and interpolated to a high resolution grid. The high resolution grids are generally incorporate digital elevation models. Low resolution GCMs does not correspond to realistic topography. . It is well known that involving or building statistical relationships between observed climate fields and the grid cells of the DEM is more complicated. There is limited ability to compensate for features in the GCM, such as orographic precipitation over mountainous regions that can be displaced geographically due to smoothed topography. hus statistical downscaling fails as it has limited ability for displaced features such as mountain ranges. Dynamical downscaling or regional climate modeling (RCM) also relies on output from GCM simulations. But provides internally consistent atmospheric parameters and true simulation capabilities. If we can compromise on relative abilities it can be said that statistical downscaling is fast while dynamic one is time consuming.

Hi Muhamad. If you are interested in downscaled data based on future climate projections, I suggest you take a look at the tools available at http://ccafs.cgiar.org/spatial-downscaling-methods#.UqXZhKWIu-Q I don't have a solid basis for saying that one of them is best. A follow-up comment suggested that you are interested in ENSO. I might be able to point you to something on this topic if you clarify what you want to do.

See http://www.pnas.org/content/early/2014/05/28/1314787111.abstract for a study of this question. The abstract:

One of the largest concerns about future climate change is its potential effect on food supply. Crop yield projections require climate inputs at higher resolution than typical for global climate models, and the computationally expensive technique of dynamical downscaling is widely used for this translation. We simulate maize yield in the United States to test whether current dynamical downscaling methods add value over simpler downscaling approaches. Our results suggest that they do not. Addressing large-scale systematic biases in climate output may be a higher priority for understanding future climate change impacts.

This paper on "Representative general circulation models selection and downscaling of climate data" might be helpful:

https://rmets.onlinelibrary.wiley.com/doi/full/10.1002/joc.6447

You may want to have more insights from this recent work detailing the different down scaling approaches