Talkin Bout the Resolution

1
Talkin Bout the Resolution

• Or
• Physically based rainfall disaggregation
• Or
• Trying to get useful forecasts of convective

2
What Im Gonna Tell You About

• Statement of the problem
• Long, roundabout approach to solving the problem,
  using
• Remote Sensing
• Simple (very) convective model
• Some results
• Everything I neglected to mention

3
THE PROBLEMNWP Model Convective Rainfall
Forecast Fields

• Color Scale

1-2mm
100 km
Rainfall
100km
4
Surface heating and rainfall
Thunderstorm tracks 22/07/96
Fox and Collier (2000)
5
Two ways to go

• We can throw away NWP model data
• This means doing our own thing
• We need to know something about water vapor
  availability
• We can keep the NWP output
• This means we develop a dimensionless
  disaggregation parameter
• We use all that lovely water budget information
  wrapped up in the model

6
Both ways are based on

• Simple estimation of convective rainfall rate
  from

R Rainfall rate ? Density of air ?s
saturated mixing ratio W vertical velocity
7
ALL ARE MEASURED AT AN ALTITUDE OF Z

• Z is the height of the cloud base
• We assume this is the top of the atmospheric
  boundary layer (potential temperature step)
• We can find Z from a simple convective model if
  we know the surface sensible heat flux

8
Finding wZ

• wZ can be found from the sensible heat flux and
  the surface temperature

9
Assuming adiabatic parcel ascent within a
boundary layer of constant potential temperature
Tz - Parcel temperature Te environment
temperature Ta surface air temperature
10
We need from remote sensing the following

• Surface temperature
• Sensible heat flux
• Depends on evapotranspiration and NDVI
• NDVI
• Evapotranspiration
• Found from change in soil moisture content

11
Remote sensing

• From AVHRR
• Surface temperature
• NDVI
• From SAR (well, sort of)
• Soil moisture
• Analysis in ERDAS Imagine

12
Features of the SHF field

• Lots of SH
• In urban areas
• High absorption of solar radiation
• Little evapotranspiration
• Not much SH
• Upland forested moors
• Lots of evapotranspiration

13
Neils theory of organized vegetation

• If vegetation is arranged in an organized pattern
  and the wind blows in the right direction then
• One can enhance local circulations (NCMC
  Non-classical mesoscale circulations)
• One can change rainfall patterns
• Need to get the right wavelength
• Need help from orography
• Need to get money to investigate!

14
Into the future

• If we can predict the required fields for a few
  days ahead we can make forecasts
• Net radiation
• We know solar elevation etc
• We have measured albedo
• NDVI is not going to change much
• SMC we model using a 1/day1/2 relationship
• IT ALL FALLS APART IF IT RAINS!

15
Case study

• NW England
• Manchester
• Mersey Basin
• West Pennine Moors
• July 1996

16
If we ignore NWP output

• We need knowledge of water vapor availability
• We can get this from remote sensing also BUT
• We cannot forecast forward on time
• Unless we use some form of frequency distribution
  of precipitable water to produce a probability of
  certain rainfall and stream flow

17
On the other hand

• If we assume that the model is pretty good at
  gross water balance and that the average
  convective fall over the grid square is accurate
  THEN
• We need a normalized parameter to distribute
  (disaggregate or downscale) the NM rainfall

18
CDP Convective Disaggregation Parameter (f)
Mean f
Normalized f
19
Disaggregated square convective rainfall