Presentation ~ above theme: "Sources the Error in NWP prediction or every the excuses You’ll ever before Need Fred Carr COMAP Symposium 00-1 Monday, 13 December 1999."— Presentation transcript:




You are watching: Which is not a source of forecast error in nwp

1 resources of Error in NWP forecasts or all the sorry You’ll ever before Need Fred Carr COMAP Symposium 00-1 Monday, 13 December 1999

*

2 development NWP has come to be an indispensable device for the forecaster, however it is important to know its limitations. There are many sources of feasible error in one NWP forecast. If you save these resources in mind together you research NWP products, you should be able to make more intelligent usage of the commodities in your forecasts. These sources of error deserve to be grouped right into three categories: A.Errors in the Initial problems It is a facility process to collect data native observations and also get them right into a type that one NWP model deserve to use. Errors can take place at several steps along the way, and also grow out of constraints of the data sources themselves.

*

3 Intrinsic Predictability constraints Errors in the Initial conditions 1Observational Data Coverage aSpatial thickness bTemporal Frequency 2Errors in the Data aInstrument Errors bRepresentativeness Errors 3Errors in Quality manage 4Errors in Objective analysis 5Errors in Data assimilation 6Missing Variables Errors in the Models 1Equations of movement Incomplete 2Errors in number Approximations aHorizontal Resolution bVertical Resolution cTime Integration Procedure 3 Boundary problems aHorizontal bVertical 4 Terrain 5 Physical processes aPrecipitation iStratiform Precipitation iiConvective Precipitation bRadiation cSurface power Balance dBoundary great iSurface great iiEkman or mixed Layer Intrinsic Predictability constraints

*

4 advent B.Errors in the design A design is by definition an approximation of reality, and also although NWP models continue to flourish in complexity, they can not take right into account all determinants that affect the weather. The numerical solution of these models by computer systems introduces added error. C.Intrinsic Predictability Limitations also with error-free observations and a “perfect” model, projection error will thrive with time. Over there is one intrinsic border to the variety of a valuable forecast. This range is short for small-scale phenomena and increases because that synoptic and planetary-scale features.

*

5 Intrinsic Predictability constraints Errors in the Initial conditions 1Observational Data Coverage aSpatial density bTemporal Frequency 2Errors in the Data aInstrument Errors bRepresentativeness Errors 3Errors in Quality control 4Errors in Objective evaluation 5Errors in Data adaptation 6Missing Variables Errors in the Models 1Equations of motion Incomplete 2Errors in numerical Approximations aHorizontal Resolution bVertical Resolution cTime Integration Procedure 3 Boundary conditions aHorizontal bVertical 4 Terrain 5 Physical procedures aPrecipitation iStratiform Precipitation iiConvective Precipitation bRadiation cSurface power Balance dBoundary class iSurface class iiEkman or combined Layer Intrinsic Predictability limitations

*

6 also with error-free observations and also a “perfect” model, estimate errors will thrive with time. No matter what resolution of monitorings is used, over there are constantly unmeasured scales of motion. The power in this scales transfers both up and also down scale. The upward transfer of power from scales less than the observing resolution represents an energy source for larger- scale motions in the environment that will not be present in the numerical model. Thus, the genuine atmosphere and the environment that is represented in the numerical version are different. For this reason, the model forecast and the real setting will diverge with time. This error expansion is approximately equal come a doubling of error every 2-3 days. Therefore, even very little initial errors can an outcome in major errors because that a long-range forecast. The problem just stated is the essence of chaos theory used to meteorology. This concept proposes the nothing is completely predictable, that also very little perturbations in a system result in unpredictable alters in time.

*

7 Intrinsic Predictability limitations This graphics illustrates the impact of intrinsic predictability constraints on estimate skill. Forecasts based upon climatology will have actually a relatively high level the error, yet will remain continuous over time. Forecasts based upon persistence (i.e., everything is happening currently will occur later) are virtually perfect in ~ extremely quick range, but quickly deteriorate. Current models carry out well at quick ranges, however eventually carry out worse than climatology. A forecast that is worse than climatology is taken into consideration useless.

*

8 even the ideal model we deserve to envision will, because that reasons just discussed, create forecasts the deteriorate in time to a quality reduced than those based on climatology. Our present forecast models have skill as much as the 5-7 day variety on the synoptic scale for 500 hPa heights. (Occasionally they have skill in ~ 15-30 days because that time-averaged planetary waves.) They show much much less skill for obtained quantities such as vorticity advection or precipitation. Intrinsic Predictability restrictions

*

9 A associated predictability limitation is the intrinsic error development will contaminate smaller scales faster than bigger scales. In other words, a small-scale phenomenon will certainly be less well estimate than a large phenomenon in the same variety forecast. However, mesoscale/convective scale predictability might not monitor this smooth progression because of its highly intermittent nature. For example, a rotating supercell thunderstorm might have more predictability (2-6 hr) 보다 an airmass thunderstorm (1 hr). Topographically and/or diurnally-forced circulations such as drylines and sea breezes are much more predictable 보다 squall lines.

*

10 finish Comment It used to be assumed that the errors as result of horizontal resolution constituted around 30% that the total forecast error. However, as result of faster computer systems (which have actually allowed much more accurate numerical schemes and higher resolution) this is no much longer the case. Currently, the largest source of error is more likely to be the unavailability of high resolution data end the whole forecast domain. One can now speak that brand-new (and accurate) observing systems, which measure the variables we need under every weather conditions, are the best way to boost NWP forecasts.

See more: How Much Does A Dishwasher Weigh, How Heavy Is A Residential Dishwasher

Improvements in computer systems (which may allow higher horizontal and vertical resolution) and in the parameterizations that physical procedures within the models will help, yet to a lesser level than brand-new observing systems.

*