Global Precipitation Measurement: Past, present, and future challenges Chris Kidd ESSIC,University of Maryland, and NASA/Goddard Space Flight Center, USA.

Global Precipitation Measurement:Past, present, and future challenges

Chris KiddESSIC,University of Maryland, and

NASA/Goddard Space Flight Center, USA

Overview

Why precipitation? - The Value of Water

- Facts and figures

Precipitation Measurement - Surface measurements

- Satellite retrievals

- Validation & inter-comparison studies

Challenges - Characterisation of precipitation

- Mapping and integration

ESSIC/UMD 6 February 2012Goddard Space

Flight Center

Background

The King’s School, Grantham (Maths, Physics, Geography)

University of Nottingham (Geography – Cartography & Earth Obs.)

University of Bristol (PMW retrievals of precipitation over land)

USRA - NASA/GSFC (inter-comparisons & merged products)

University of Birmingham (Satellite meteorology and climatology)

ESSIC - NASA/GSFC (multi-source/scale precipitation products)


Flight Center

Why precipitation?

“Our knowledge of the time and space distributions of rainfall, soil moisture, ground water recharge, and evapotranspiration are remarkably inadequate, in part because historical data bases are point measurements from which we have attempted extrapolation to large-scale fields.”

P.299 National Research Council (1991) Opportunities in the hydrologic sciences, National Academy Press

“…critical atmospheric variables not adequately measured by current or planned systems [include] precipitation.”

UK Met Office. NERC CEOI Workshop, 13/11/09

Precipitation is ultimately the input for all hydrological systems

Personally – it combines geography & weather


Flight Center

Precipitation – the enigmaQuantifying precipitation, its accuracies and errors is

extremely problematic; critical issues affecting and influencing the observation and measurement of precipitation are:

i) the characteristics of the phenomenon being observed;ii) the observational capability of the sensor;iii) the interpretation of the observations and the derived

parameters, and;iv) the perceived versus real requirements of the subsequent

applications.

Defining the accuracy and associated errors of any precipitation observation or measurement is therefore a multidimensional and inexact problem.


Flight Center

Water factoids

1 mm per square metre = 1 litre (or 1 kg)1 mm per square kilometre = 1,000,000 litres or 1000 tonnes

so, D.C. has ~1000 mm/yr ≡ 1,000,000 tonnes/yr/km2

“More than 2.8 billion people in 48 countries will face water stress or scarcity conditions by 2025.” WaterFootprint.org & WWF

Currently fresh water costs ~$2 per cubic metre, globally precipitation ‘contributes’ $258 trillion annually.

Over the US alone, precipitation is ‘worth’ $13 trillion annually.


Flight Center

Water – facts & figures


Flight Center

UK Lake District floods 2008 & 2009

2008 floods

2009 floods

Surface measurements

Clee Hill radars (C-band vs ATC)

Micro rain radar0.2 mm/tip ARG100 gauge 0.1mm/tip Young’s Gauge

Conventional Observations

20,000Raingauges

Radar duplicates rain-gauge coverage

Precipitation is highly variable both temporally and spatially.

Measurements need to be representative


Flight Center

Variance explained by nearest station

Jürgen Grieser

NOTE: Monthly data: shorter periods = lower explained variance


Flight Center

What is truth? Co-located 8 gauges / 4 MRRs


Flight Center

Precipitation accumulation

0.0

2.5

5.0

7.5

10.0

12.5

15.0

1 6 11 16 21 26 31 36 41 46 51 56

Time (Minutes after 16:00, 06/07/06)

Ac

cu

mu

lati

on

(m

m)

12345678MRR1MRR2MRR3MRR4


Flight Center

Surface measurements summary

Representativeness of surface measurements:• Over land generally good, but variable• Over oceans: virtually none-existent

Measurement issues:• Physical collection – interferes with measurement (e.g.

wind effects – frozen precip, etc)• Radar – imprecise backscatter:rainfall relationship (also

clutter, range effects, bright band, etc)

Satellites offer consistent, regular measurements,global coverage, real-time delivery of data


Flight Center

Satellite precipitation observation capabilities

1960

1970

1980

1990

2000

2010

2020

1959 Vanguard 21960 TIROS-1

1966 ATS-1

1983 NOAA-8

1987 SSM/I

2014 GPM

2011 Megha-Tropiques

?

1974 SMS-1

1978 SMMR

Visible

Active MW

2003 SSM/IS2002 MSG

2006 Cloudsat

1998 AMSU1997 TRMM

Passive Microwave

Infrared

1988 WetNet

1990 PIP-1 1993 PIP-2 1996 PIP-3

1989 AIP-1 1991 AIP-2 1994 AIP-3

2001 IPWG2004 PEHRPP


Flight Center

Satellite retrieval of precipitation

Note: Observations are not direct measurements

Visible (including near IR)• Reflectance, cloud top properties (size,

phase)

Infrared• Thermal emission – cloud top

temperatures → height

Passive Microwave• Natural emissions from surface and

precipitation (emission and scattering)

Active Microwave• Backscatter from precipitation particles


Flight Center

Combined Vis/IR & microwave techniques

Rationale: Observations more directly related to hydrometeors

Rationale: Observation of cloud top properties (temperature/size), but indirect

Observations: Frequent observations (30mins); Good spatial resolution (1-4 km)

Observations: Infrequent observations (2/sat/day); Poor spatial resolution (5-25 km)

Combine directness of MW observations with the resolution/frequency of IR observations

Vis/IR Microwave (active/passive)

Calibration of Vis/IR-derived properties with microwave

observations

Advect microwave estimates with information from IR

observations

☺☺

☺

☺

+ model information....


Flight Center

‘Global’ <30 minute

<12km rainfall

estimatespossible

Infrared daily

estimate

Passivemicrowave

daily estimate

RegionallyCalibrated

product

PM-IR products

Can we generate

1km, 1min global

estimates?


Flight Center

Advection/Morphing products

Wind vectors derived from MSG 15 minutes data

(simple correlation match)PMW estimates advected using MSG

wind vectors: 0745-0930

Basis of ‘CMORPH’ and GSMaP techniquesuses forwards and backward propagation of PM rainfall

12 May 2003MSG – SSMI

study


Flight Center

Precipitation product inter-comparisons


Flight Center

NASA WetNet: Tallahassee c.1989


Flight Center

NASA WetNet PIP-1 Bristol c.1991


Flight Center

GPCP AIP-3 Shinfield Park c.1993


Flight Center

IPWG#4 CMA Beijing 2008


Flight Center

International Precipitation Working GroupNear real-time inter-comparison of model & satellite estimates vs radar/gauge


Flight Center

IPWG European Inter-comparison


Flight Center

IPWG: European region 07/11-01/12

July August September October November December January

Satellites ~same as models in summer; models better in winter

Cor

rela

tion


Flight Center

Radar vs gauge dataR

adar

(da

ily in

tegr

ated

)G

auge

dat

a


Flight Center

Statistics: blame it on the weather!

Movement:Is the movement perpendicular or along the rain band?

IntensityWhat is the range of values within the rain area?

Size/variabilityWhat is the size and variability of the rain area(s)?

Statistical success has as much to do with meteorology as the algorithms ability…

Sensor field-of-view

Type of cloud/

rain


Flight Center

3-hourly/0.25 degree data availability

NOTE: not all ‘data’ is real ‘data’


Flight Center

Precipitation Products, Europe 2009

0 1 2 3 4 5 6 7 8 9 10mm/day

MWCOMB CMORPH ECMWF

PERSIANN 3B42RT GPCC gauge

Orography & high latitudes still presents a challenge to retrieval techniques

2009 Annual Mean


Flight Center

Precipitation Products, Africa 2009

GPCC gauge

ECMWF

3B42RT

NRLBLD CMORPH

PERSIANN

Over central Africa: PMW overestimates (convective); gauges underestimate (representativeness); model ~right?


Flight Center

2009

Ann

ual M

ean

109876543210

mmd-1

Retrieval challenges

- Good agreement in Tropics- Poorer in the extra-Tropics

0

1

2

3

4

5

6

7

60 50 40 30 20 10 0 -10 -20 -30 -40 -50 -60

Latitude

Mea

n a

nn

ual

rai

nfa

ll (m

m/d

ay)

3B42RT

CMORPH

CPCMMW

GPCC05

NRLBLD

PERSIANN

ECMWF

Land-only 20W-20E latitudinal profile

60°N-60°Slimit max.

Extra-Tropical

30% satellite :gauge difference

Agusti-Panareda and Beljaars (2008)


Flight Center

Correlation Bias-ratio2005-11


Flight Center

SE England analysis (vs radar)

impr

oving

per

form

ance

Timelineposition

evaluation ofindividual 0.25°x 0.25° boxes


Flight Center

Dirunal Cycle UK

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Probability of Detection

Fa

lse

Ala

rm R

ati

o 3B42RT

CMORPH

CPCMMW

PERSIANN

ECMWF

GPI

NRLBLD

Dirunal Cycle DE

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Probability of Detection

Fa

lse

Ala

rm R

ati

o 3B42RT

CMORPH

CPCMMW

PERSIANN

ECMWF

GPI

NRLBLD

09-12

12-15

12-15

09-12

12-15

12-15

ECMWF: evident diurnal cycle in performance

CMORPH: over Germany performance in JJA ≈ that of ECMWF

Diurnal statistical performance (JJA)

Temporal/spatial analysis can help identify surface/satellite errors more easily

Generated from 3-hourly accumulations


Flight Center

Future Challenges I

Characterisation of precipitation

High latitude retrievals• Light precipitation• Snowfall and mixed-phase precipitation• Land/ocean/coastline consistency

The retrieval of precipitation at higher latitudes is more challenging due to the physically diverse nature of the weather systems and backgrounds.


Flight Center

High-latitude processes

Cryospheric processes are complex with longer time-scale water cycle implications than the Tropics


Flight Center

High latitude precipitation

Validation instrumentation at high latitudes to observe and

measure precipitation


Flight Center

Model vs satelliteE

CM

WF

3B42

RT

3-hourly precipitation accumulations for 1 June 2007

Clear differences between identification (or definition) of precipitation


Flight Center

Distribution of light precipitation• Light rainfall becomes increasingly important towards the polar regions

• COADS data shows light precipitation occurrence >80%; ~50% in mid-latitudes

• European radar suggests ~85% of precipitation <1 mmh-1 (35%<0.1 mmh-1)

• Accumulation of light precipitation is smaller, particularly in the Tropics

Current satellite techniques do not retrieve light precipitation well


Flight Center

LPVEx: 21 September 2010

Aranda

Jarvenpaa

Rainrate

Rainrate

AMSRV10 rain10:23Z

Significant coastline problems

and light-rain detection.

MRR data


Flight Center

LPVEx: 14 October 2010

Jarvenpaa

AMSRV10 rain10:29Z

Rainrate

MRR data


Flight Center

Western Europe 5-6 February 2012

SSMIS F17 0557Z 6 February 2012

Fallen snow and falling snow do not necessarily have unique signatures

surface snow

falling snowfalling snow

surface snow

SSMIS F17 0610Z 5 February 2012

H150 183±1 183±3 183±6

H150 183±1 183±3 183±6

Future Challenges II

Mapping & integration of data

• Representation/mapping of data globally - in particular beyond the local areas or outside the tropics

• Utilisation of multi-scale data sets – how to integrate different data sets to improve products

• Sub-pixel resolution requirements – migration from coarse-resolution to fine-resolution products (both for actual precipitation products & processing)


Flight Center

Mapping

How should data be mapped?

Should data be mapped at all?

‘Standard’ mapping for global precipitation is the lat/lon (CED) grid: advantages include simplicity, ease of use and interpretation, but the main disadvantage is the non-equal area nature of the mapping, particularly at higher latitudes.

Critical for regions outside the tropics: at 60°N/S the E-W distance is 0.5 that of the N-S distance


Flight Center

75N 0.259

60N 0.500

45N 0.707

30N 0.866

Equator 1.000

30S 0.866

45S 0.707

60S 0.500

75S 0.259

Scale relative to the Equator (=1.00)

Local area mapping

Rationale: polar coverage and finer resolutions necessitate the generation of products of equal area.

Method: provide local-area mapping of products.- this should not take any more processing time- observations are mapped to each tile (via look-up-table)- each region has a small overlap with the neighbouring

tile to allow consistency- motion vectors and products are generated for each tile- products are saved as lat/lon/value.


Flight Center

Local area mapping errors

0 5 50 km error


Flight Center

Advection techniques

CMORPH motion vectors, 2.5° resolution

Resolution

Can we adequately resolve the peculiarities associated with the movement of precipitation?

At present we only have very crude representations of precipitation motion.


Flight Center

High resolution climatologiesTRMM PR data: 13 years (1997→) at ~5 km resolution.

Occ

urr

ence

of

rain

fall

Rainfall shows significant local variability linked with relief.

An

nu

al t

ota

l rai

nfa

llHimalayas

Western G

hats


Flight Center

Conclusions

Sound foundations;• 50+ years since the dawn of the satellite era• 33 years since usable observations for precipitation• 15 years since ‘first’ precipitation mission

Significant progress in precipitation retrievals;• coarse resolution through to ~0.25° 3-hourly estimates• - although generally limited to 60°N-60°S

Current & future issues;• High latitudes to complete global precipitation estimates• Fine resolution, multi-source retrievals - globally


Flight Center

ECMWF operational model, annual precipitation 2009

The ultimate goal...


Flight Center

Global Precipitation Measurement: Past, present, and future challenges Chris Kidd ESSIC,University of Maryland, and NASA/Goddard Space Flight Center, USA.

Documents