Quantifying Global Oceanic Precipitation by Combined Use of In Situ and Satellite Observations P' Xi

1
Quantifying Global Oceanic Precipitationby
Combined Use of In Situ and Satellite
ObservationsP. Xie, R. Joyce, J.E. Janowiak,
and P.A. Arkin
2
Objective

• To review the current status of constructing
  observation-based data sets of global oceanic
  precipitation
• To provide suggestions on what we need to do to
  improve the quantitative documentation and
  monitoring of oceanic precipitation

3
Combination of In Situ Satellite Obs Used in
Defining Precip. Analysis

• Satellite observations provide information of
  spatial / temporal variations
• In situ instruments (ships, buoys, atolls..) make
  direct measurements
• Merging improves quality of oceanic precip
  analysis
• Various merged / combined analyses (e.g. GPCP,
  CMAP, TRMM) present similar spatio-temporal
  variation patterns

4
Problems in Existing Precip Data Sets

• Uncertainty
• in quantitative magnitude
• Inhomogeneity
• in Long-term time series
• Poor Quality
• over high latitudes
• Coarse Resolution
• in long-term data sets

5
Quantitative Uncertainty 1Differences among
Data Sets

• Three sets of observation data sets used
• CMAP / GPCP / TRMM
• Annual climatology (mm/day) for 1988 2000
• Largest uncertainties (standard deviation among
  observations) over ITCZ and high latitudes
• Standard Deviation about 10 of the mean values

6
Quantitative Uncertainty 2Sources of the
Differences

• Input Satellite Estimates
• Satellite obs. (IR, PMW)
• Retrieval algorithms
• Calibration methods/Data
• SSM/I-based precip from two different algorithms
• Bias Adjustment Methods
• Against one satellite estimates (e.g. GPCP)
• Against in situ data (e.g. CMAP)

7
Long-Term Inhomogeneity 1Differences over the
Data Period

• Inhomogeneity observed in many long-term data
• Rotated EOF of GPCP monthly anomaly for 1979
  2005
• Mode 6 associated with inhomogeneity associated
  with the use of OLR-based precipitation estimates
  before

8
Long-Term Inhomogeneity 2Sources of the
Differences

• Input Satellite Data
• MW not available before 1987
• Differences between IR-MW
• ? Histograms of IR- MW-based monthly precip
• Satellite orbit changes
• (observing different phases of a diurnal cycle)
• Instrument calibration

9
Poor High-Latitude Estimation Problems and Causes

• PMW unable to detect precip over icy surface
• PMW estimates over open ocean miss light precip
• IR-based estimates relate precip to cloudiness
• ? SSM/I PMW estimates of Wilheit et al.

10
Spatial / Temporal Resolution Long-term data
sets vs State-of-the-art estimates for recent
period

• Coarse spatial / temporal resolution for
  long-term data sets
• 2.5olat/lon
• monthly / pentad
• Fine-res new satellite estimates too short to
  define climatology

11
Critical Elements Need to be Addressed for
improved observation of oceanic precipitation

• In Situ Measurements
• buoys, ships, special field experiments ..
• Satellite Estimates
• new instruments, new technology,
• new networks
• Combining Information from Various Sources
• different satellites
• in situ satellites
• precip other parameters (e.g. moisture,
  temperature ..)

12
In Situ Measurements

• Direct measurements
• Calibration and assessments of satellite
  estimates
• Correction of local bias in satellite data
• ? Comparison of three SSM/I-based estimates
  with buoy
• What we need for future improvements
• Quantitative accuracy (wind correction)
• Extended buoy networks over extra-tropical oceans
• (esp. storm tracks)

13
Satellite Estimates

• Quasi-complete spatial coverage
• Regionally / temporally varying systematic error
• Poor quality over high latitudes
• Inhomogeneity in long-term time series
• Things Under Going
• Global Precipitation Measurement (GPM)
• Improving estimation of high-latitude precip
  using data from AMSU data
• ? Preliminary results from an MIT group

Histogram of Precipitation
Courtesy of C. Surussavadee and D. Staelin
14
Combining Information from Multiple Satellites

• Defining precipitation estimates with improved
  quantitative accuracy at a fine resolution
• CMORPH stands out as the best products of hi-res
  precipitation for recent years
• CPC is working on the further refinement of
  CMORPH through using Kalman Filtering technique
  and including inputs from more satellites

15
Combining Information from Different
Observational Platforms

• Removing of local bias in satellite estimates
  requires input of information from in situ
  observations
• Experiments with gauge / satellite data over
  China demonstrate successful correction of biases
  and improvements in patterns

16
Conclusions / Recommendations

• Substantial progress made over the past decade in
  documenting seasonal cycle, interannual
  variations intraseasonal variability of oceanic
  precipitation
• Problems exist in current data sets in
  quantitative uncertainty, long-term
  inhomogeneity, high-latitude estimation quality,
  resolution
• Users requirements need to be identified and
  weighted to set goals for the next decade
• Combining in situ and satellite information an
  effective way to construct oceanic precipitation
  before data assimilation ultimately outperforms
• We need to study and discuss how we, in
  collaborations with other communities, can
  improve the observations of oceanic precipitation
  (in situ, satellites, combining)

17
North Pole Precipitation July 20-21, 2006
AMSU-derived precipitation over North Pole sea
ice (pink) evolution over 24 hours High surface
elevation is problematic (dark pink)
Courtesy of C. Surussavadee and D. Staelin