Impacts of high-resolution Himawari-8 AMVs on TC forecast in - PowerPoint Presentation

Slide1

Impacts of high-resolution Himawari-8 AMVs on TC forecast in HWRF

Masahiro Sawada (Dr)

Visiting Scientist from

Meteorological Research

Institute, JMA

Acknowledgements

: HWRF group at EMC & DTC

Slide2

contents

Background & Objectives of this studyWhat is high-resolution Himawari-8 AMV?Model & Experimental designOutline of target cases

ResultsDetailed analysisSummary2

Slide3

TC f

orecast skill at JMA a few years ago…

Although track forecast errors have decreased to less than half of those obtained 30 years…the intensity forecast errors exhibit an

increasing behavior. Annual mean official TC forecast errors in RSMC

Tokyo (Ito 2016)

TrackPmin

Vmax

Slide4

Introduction & background

Accurately estimating the tropical cyclone (TC) intensity and structure is essential to improve TC forecast and to diagnose numerical model characteristics. Several previous studies show the assimilation of atmospheric motion vectors (AMVs) derived from satellite images benefits TC forecast using WRF/HWRF (Wu et al. 2014, 2015,

Velden et al. 2017, Zhang et al. 2018).

However, data assimilation has not been applied in the WPA basin yet. The main reason is observations around the TC are sparser in the WPA compared to the ATL and

EPA.

Slide5

Introduction & background (continued)

Satellite observations, or satellite retrieved products such as AMVs are available globally and over the ocean without (temporal) interruption!Himawari-8

provides high spatiotemporal resolution AMVs and it has a potential to better capture the TC structure and initial conditions of the surrounding environment, leading to more skillful forecasts.

Can we simply expect the high-resolution AMVs fill the sparse observations over the ocean, and is data assimilation in HWRF applicable for WPA basin?

Slide6

Purpose of this study

Assess the capability of data assimilation for the WPA

TCsInvestigate what

impacts the high resolution Himawari-8 AMVs have on the

TC forecastsUnderstand

how impacts ariseExplore

better ways to make more effective use of the Himawari-8 AMVs by changing the assimilation method

- TC

surrounding flow

and

track forecast

are improved

.

- DA

in high-resolution will provide better TC structure, reducing intensity forecast error.

Expectation

Slide7

Brief description of Himawari-8

Himawari-8 is next generation geostationary satellite (same as GOES-16).

Himawari-8 became operational at JMA in July 2015.16 channels, 10-min full disk scan, 2.5-min targeted region.AMVs are derived from full disk & targeted region scans. AMVs of full disk is routinely provided hourly.

In this study, AMVs of full disk are used for assimilation experiments because it is provided without interruption and can be used to monitor more than two TCs at the same time.

(From JMA webpage)

Slide8

What is high-resolution Himawari-8 AMVs?

Current Himawari-8 AMVs~0.3x0.3 deg, hourly

High-res Himawari-8 AMVs (H8AMV) ~0.2x0.2 deg,

30-min (Experimental product of MSC/JMA)

GSI

GSI-AMV

GSI-AMV2Assimilated Himawari-8 AMV for D02: examples

Slide9

Model & experimental design

Name

Description

CTL

cycle experiment without DA (vortex initialization only)GSI

cycle experiment with 3DVAR

using the same dataset as GFS

GSI-AMV

GSI+

H8AMV (add AMVs between -1~+1 h from analysis time)

GSI-AMV2

GSI+

H8AMV (add AMVs between -3~+1 h from analysis time)

Model: NCEP Operational HWRF (2017 Operational configuration)

except for POM ocean coupling (for 2017 HWRF, HYCOM is used for WP)

Horizontal resolution: 18, 6, 2km for D01, D02, and D03 respectively

Vertical layer: 61 up to 10

hPa

Initial condition: GFS analysis for D01 and vortex initialization

126 hour forecast

for

each

cycle

72

cycles per one

experiments

Slide10

Outline of

targeted Typhoons

Nepartak

(02W

) Meranty(16W)

Megi(20W)70kt/day!Extreme RI

Track & intensity data from JTWC Besttrack

70kt/day!

Extreme RI

30kt/day

2016

Slide11

Results

Track forecastTrack verification –all cases–Track verification –each case–Intensity forecast

Intensity verification –all cases–Intensity verification –each case–Size forecastsize

verification –all cases–11

Slide12

Track

verification -all cases-

- GSI-AMV2 provides best track forecast for F84-126, followed by GSI-AMV. => The more H8AMV are, the better track forecast skill becomes.- GSI was the worst for F66-120 in 4 experiments. => Only DA in high resolution grid space cannot improve track forecast.

Track error (km)

Slide13

Track

verification -each case-

- Positive impact of H8AMV on track comes from Nepartak case.- Degradation can be seen in the Meranti case, unfortunately.

- Neutral for Megi case.- Case to case variability.- Note that track errors of Meranti & Megi cases are quite smaller than that of Nepartak case.

NepartakMeranti

MegiCTL

GSIGSI-AMVGSI-AMV2

Slide14

Intensity verification -all cases-

- All experiments were too weak because extreme RI was not captured.- GSI-AMV2 was the largest Intensity errors in all experiments for shorter lead time.

- This would be caused by larger weak intensity bias.HWRFGSI

AMVAMV2

RMSE

BIAS

Slide15

Intensity verification -each case-

Nepartak

Meranti

MegiHWRFGSI

GSI-AMVGSI-AMV2

<= GSI-AMV2 has smallest errors!

<= GSI-AMV2 has largest errors…

Case-to-case variability is larger in intensity than in track forecasts

Slide16

Size verification (BIAS) -all cases-

- It is hard to see differences between experiments, meaning no significant impacts.- In the R34, 50, & 64, the sizes with GSI-AMV2 tend to be small for shorter lead time compared to CTL.

HWRFGSIAMV

AMV2HWRF

GSIAMVAMV2

R34

R50

R64

RMW

Slide17

What processes account for impacts?17

High-resolution AMVs =>

Improve track forecast skill, but degrade intensity forecast skill…why?

Diagnosis of forecasted track error based on optimal steering flows -Nepartak case-Comparison of initial vertical structures of axisymmetric mean composite -Meranti case-

Slide18

Track error diagnosis –method–18

Estimate optimal steering radius

& depth according to Galarneau & Davis (2013) to identify cause of track error differences. Cyclonic circulation associated

with TC was removed. Steering flow is calculated from 13 different vertical depth (50−650 hPa with

a fixed base at 850 hPa), and 8 different radii (100−800km).

The radius & depth combination that produces the smallest vector difference between the averaged steering flow and the actual track motion is chosen as the optimal steering flow radius & depth.

Slide19

Track error diagnosis –TC structure–

19

Results of optimal steering flow diagnosis for Nepartak case.No systematic difference in optimal steering radius and depth...

=> Reduction of TC track error is not caused by the TC structures through beta-gyre effect.=>How about synoptic flows?

Depth (hPa)Radius (km)

For the most of time, 300~400kmFor the most of time, 400~500hPa(850-450hPa ~ 850-350hPa)

Optimal steering radius

Optimal steering depth

CTL

GSI-AMV2

Radius(km)

317

314

Depth(

hPa

)

482

492

Slide20

Track error diagnosis –Synoptic flow–

20Steering flow profile difference between GSI-AMV2 & CTL.Steering

flow differences became remarkable with time from upper to lower troposphere.=> Synoptic flow difference is more likely to reduce the track error biases, not caused by TC structure (size & depth) differences.

Vector difference is largeColor:

Vector Difference in steering flows (m/s)Arrows: Difference in steering vectors calculated within

300-km radius of TC center at each forecast lead time. Height (hPa

)

Slide21

Weak intensity bias

HWRF

GSIGSI-AMVGSI-AMV2

Why

was intensity forecast with GSI-AMV2

degraded for shorter lead time?

BIAS of Vmax (

kts

)

Slide22

Vertical structure difference in Vt

With DA, initial

vortices became stronger outside RMW.=> larger inertial stability.=> higher resistance of radial displacement.

=> would weaken low-level inflow.Composite of vertical Structure axisymmetric mean

Vt at F00.Normalized radius by RMW at 2-km height is used.

Cycle averaged for

3-20 of case 16W.Color: difference between each experiment minus

CTL.

22

TC center

outer

Slide23

Vertical structure difference in low-level inflow

Inclusion

of H8AMV actually weakens low-level inflows.

=> slower intensification because of reduction of inward angular momentum transport.

Composite of vertical Structure axisymmetric mean Vr

at F0-24.

Cycle averaged for 3-20.

Color: difference between each experiment minus HWRF.

23

Slide24

Vertical structure difference in axisymmetricity

Axisymmetricity

is defined:

Inclusion

of H8AMV enhances asymmetric structure. =>

slower axisymmetrization.

Composite of vertical Structure axisymmetric mean Vt

at F00.

Cycle a

veraged for

3-20.

Color: difference between each experiment minus

CTL.

24

(Miyamoto &

Takemi

2013)

Slide25

Vertical structure difference in RH25

Emanuel & Zhang (2017) showed intensity change

is strongly influenced by inner core moisture, seems to be related to negative bias.

Initial vortices in

GSI-AMV2 became drier within inner core region =>slower intensification(Initial vortices in GSI became drier within RMW & at upper troposphere.)

Composite of vertical Structure axisymmetric mean

RH at F00.Cycle averaged for

3-20

.

Color: difference in RH between each experiment minus

CTL.

Contour:

RH

Slide26

Brief summary of factors explaining impacts26

High-resolution AMVs =>

Improve track forecast skill, but degrade intensity forecast

skill.1. Improvement of track forecast skill=> Improvement in steering flow, not in TC structure itself2. Degradation of intensity forecast skill=> Weakening of low-level inflows, decrease in axisymmetricity,

and drying within the inner core region

Slide27

How can we make effective use of H8AMV?27

How about changing assimilation method, such as hybrid DA instead of 3DVAR?

=>Background error covariance are created from HWRF, so flow-dependent structure & cross-variable correlation would be represented properly.

Slide28

Model & experimental design

Name

Description

CTL

cycle experiment without data assimilationGSI

cycle experiment with 3DVAR

using GSI

HGSI

cycle experiment with hybrid DA which uses background error covariance created from HWRF

GSI_AMV2

3D + high-resolution

AMV derived from

Himawari-8

HGSI_AMV2

Hybrid

DA

+ high-resolution

AMV derived from

Himawari-8

Model: NCEP Operational HWRF (2017 Operational configuration)

except for POM ocean coupling (for 2017 HWRF, HYCOM is used for WP)

Horizontal resolution: 18, 6, 2km for D01, D02, and D03 respectively

Vertical layer: 61 up to 10

hPa

Initial condition: GFS analysis for D01 and vortex initialization

126 hour forecast for each cycle

26 cycles per one experiments from 2016.9.8.12 - 14.18 UTC

28

Slide29

Track forecast verification

HGSI_AMV2 provides the best track forecast for F42-66 &

F108-126.

The track error was smaller in HGSI_AMV2 than in GSI_AMV2.29

Slide30

Intensity forecast verification

The intensity error was significantly smaller in HGSI_AMV2 than in GSI_AMV2.

By applying hybrid DA, weak bias or slow intensification was greatly reduced!30

HWRF

GSIHGSIGSI-AMV2HGSI-AMV2

Slide31

Intensity forecast composite

GSI_AMV2

HGSI_AMV2

By using hybrid DA, slow intensification was alleviated.

Still could

not capture maximum intensity &

RI.

31

Black line: Besttrack

Colored line: each cycle run

Slide32

Why was intensity bias mitigated?32

High-resolution AMVs (+ 3DVAR)

=> Improve track forecast skill, but degrade intensity forecast skill.High-resolution AMVs + hybrid DA

=> Improve track forecast skill, WITHOUT degradation of intensity forecast! Why?

Slide33

Vertical structure difference in

Vt

In HGSI-AMV2, the increase of tangential wind outside RMW was reduced compared to GSI-AMV2.=>alleviate weakening of low-level inflow.

Composite of vertical Structure axisymmetric mean

Vt at F00.Cycle averaged for

3-20 of case 16W.Color: difference between each experiment minus

CTL.33

Normalized radius

from TC center (km)

Slide34

Vertical structure difference in low-level inflow

Weakening of low-level inflow was alleviated in HGSI-AMV2. =>slow intensification bias is improved.

Composite of vertical Structure axisymmetric mean

Vr at F00.

Cycle averaged for 3-20.

Color: difference between each experiment minus CTL.

34

Normalized radius

from TC center (km)

Slide35

Vertical structure difference in axisymmetricity

Enhancement of asymmetric component by inclusion

of

H8AMV was reduced using hybrid-DA.=>slow intensification bias is improved.

Composite of vertical Structure axisymmetric mean

Axisymmetricity at F00.Cycle averaged for

3-20.Color: difference between each experiment minus

CTL.

35

Normalized radius

from TC center (km)

Slide36

Vertical structure difference in RH36

Drying tendency at the inner core region was alleviated in HGSI-AMV2.

=>reduce slow intensification bias.

Composite of vertical Structure axisymmetric mean RH at F00.

Cycle averaged for 3-20

.Color: difference in RH between each experiment minus CTL.

Contour: RH

Normalized radius

from TC center (km)

Slide37

Summary

Assimilation of high-resolution AMV by Himawari-8 will be

beneficial for track forecast but it may increase

weak bias of intensity forecast.

Through the diagnosis of optimal steering flows (depth and radius), reduction of track error is caused by improvement

in the environmental flows, not in the representation of TC structure

.

Composite of axisymmetric mean structures indicates three factors related to the weak intensity bias at shorter lead time:

Weakening of low-level inflows due to high inertial stability outside RMW.

Decrease in axisymmetricity at the upper layer & outside RMW.

Drying within inner core region.

37

Slide38

Summary (continued) and future work

By combining AMV with hybrid DA, track forecast could be improved

without the degradation of intensity forecast.

Hybrid-DA reduces the weakening of low-level inflows, enhancement of asymmetric component, and drying within the inner core region. These three factors contribute to improvement of weak intensity bias.

38

Future workMore forecast cycles for robust results

Shorter assimilation cycle (hourly or less)

Capability of targeted region (or rapid-scan) AMVs

Slide39

Further experiments

Current operation HWRF uses hybrid-DA with background error covariance

(Be) created from GEFS.=> What degree is different in Be between GEFS & HWRF?Rapid-scan AMVs are really denser than AMVs of full-disk scan. But too much?

=> Are rapid-scan AMVs really useful?How about removing vortex initialization?

How about increasing # of outer loop in GSI?How about using FGAT with hourly first guesses?

Slide40

Experimental design

Name

Description

CTL

cycle experiment without DA (vortex initialization only)gHGSI

cycle experiment with hybrid-DA

which uses background error covariance created from

GEFS

using the same dataset as GFS

gHGSI-AMV2

gHGSI+H8AMV

gHGSI-RAMV2

gHGSI+H8AMV

+RSAMV

HGSI

cycle experiment with hybrid-DA

which uses background error covariance created from

HWRF

using the same dataset as GFS

GSI-AMV2

GSI+

H8AMV (add AMVs between -3~+1 h from analysis time)

HGSI-AMV2

GSI-AMV2

with

hybrid-DA

which uses background error covariance created from

HWRF

HGSI-RAMV2

HGSI-AMV2

+ RS

AMV

Slide41

Track

verification -all cases-

- HGSI-RAMV2 provides best track forecast for F72-126, followed by GSI-AMV2.

=> The more H8AMV are, the better track forecast skill becomes.- gHGSI was the worst for longer lead time in 4 experiments.

=> Only DA in high resolution grid space cannot improve track forecast.Track error (km)

Slide42

Track

verification -each case-

- Positive impact of H8AMV on track comes from Nepartak case.

- Degradation can be seen in the Meranti case, unfortunately.- Neutral for Megi case.- Case to case variability.- Note that track errors of Meranti & Megi cases are quite smaller than that of Nepartak case.

Nepartak

Meranti Megi

CTLgHGSIHGSI

GSI-AMV2

HGSI-AMV2

HGSI-RAMV2

Slide43

Intensity verification -all cases-

- All experiments were too

weak because extreme RI was not captured.

- HGSI-RAMV2 was

the smallest Intensity bias in all experiments for almost all lead time.- HGSI-RAMV2 was smaller intensity error than that some experiments. Not bad…- gHGSI-RAMV2 was the largest intensity bias for shorter lead time.

RMSE

BIAS

Slide44

Intensity verification -each case-

Nepartak

Meranti

Megi

GSI-AMV2 has smallest errors!GSI-AMV2 has largest errors…

CTLgHGSI

HGSI

GSI-AMV2

HGSI-AMV2

HGSI-RAMV2

Slide45

Size verification (BIAS) -all cases-

- It is hard to see differences between experiments, meaning no significant impacts

.

R34

R50

R64

RMW

Slide46

Slide47

Intensity verification -each case-

Nepartak

Meranti

Megi

HWRFGSIGSI-AMV

GSI-AMV2

GSI-AMV2 has smallest errors!

GSI-AMV2 has largest errors…

Case-to-case variability is larger in intensity than in track forecasts

Slide48

Size verification -3cases-

- All RMW Size forecast seems to be positive (large) and biases increases with time. - R34 biases are smaller for early forecast lead time and increases with time.

- Bias of RMW is larger that that of R34, for me it feels strange because RMW is smaller than R34 but…

RMW

R34

Slide49

Intensity verification (BIAS) –each cases-

- Largest negative bias can be seen in the

Meranti

case, especially AMV2.

- Negative biases are peak between F54-78 for both Nepartak & Meranti

cases because all experiments failed to capture ‘extreme’ RI.Nepartak

(26)Meranti(26)

Megi

(21)

Slide50

Nepartak02W

forecast

intensity composite

(c)

GSI

(b)GSI

(a)HWRF

(d)H8AMV2

- B216 is too strong in the early

intensification

stage.

- HWRF & HGSI look better than

B216 in the early

intensification

stage. GSI looks slightly better.

- However, a

ll models are too weak in the mature stage and missed RI.

(c)H8AMV

Slide51

Along-/cross-track biases

51

Track biases: GSI-AMV2 < GSI-AMV < GSI ~ HWRF=>The more H8AMV were assimilated, the smaller track biases became. 1: synoptic flow, 2: storm

structure through beta-gyre effect.

Mean track errors in the along-/cross-track directions. Dots: mean displacement of storm center from besttrack with a 24 h interval

.All cases (# 72) are averaged.Cross-error bias is significant.

Slide52

Meranti16W

forecast

intensity composite

(c)GSI(b)HWRF

(a)B216

(d)HGSI- B216 is too strong in the early intensification

stage. - HWRF & HGSI look better than B216 in the early intensification stage. GSI looks slightly better. - However, a

ll models are too weak in the mature stage and missed RI.

Replacement is needed!!!

Slide53

Why did tracks changed?

Track error difference

=

H8AMV2

- HWRF

Nepartak (km)

2016070300

HWRF

H8AMV2

2016070400

HWRF

H8AMV2

2016070218

HWRF

H8AMV2

2016070318

HWRF

H8AMV2

Q: How can we diagnose the factor which causes track change?

Meranti

Megi

Synoptic forecast skill: Z500, steering flows…

Beta gyre effect via TC size

Slide54

Forecasted Track biases

54

Mean track errors in the

lon/lat directions. Northward track error bias is significant.Track biases: AMV2 < AMV < GSI ~ HWRF

=>The more Himawari-8 AMVs were assimilated, the fewer track biases became. 1. synoptic flow. 2. storm structure through beta-gyre effect.

Longitude track bias (km)

Northward bias

Southward bias

Latitude track bias (km)

HWRF

GSI

AM

V

AMV2

Dots: mean displacement of storm center from besttrack with a 24 h interval.