Variation of daily average rainfall from June 2014 to May 2015 over Bangladesh measured by IMERG and rain gauge. Scatter plots between average daily rainfall between IMERG and rain gauge from June 2014 to May 2015. Insufficient rain gauge networks throughout the country sometimes provide incomplete information on the distribution of rainfall.
The daily, monthly, and seasonal correlation coefficients (CC) between IMERG and rain gauge data are 0.99. Inadequate rain gauge networks across the country sometimes provide incomplete information on rainfall distribution (Islam and Uyeda 2005). Examining daily, monthly and seasonal variations in IMERG rainfall and BMD rain gauge data in Bangladesh.
To study the thresholds and different percentiles of precipitation events measured by IMERG and rain gauge. To find out the Root Mean Square Error, correlation coefficient and standard deviation between IMERG and rain gauge data.
Review
2016) studied the similarity and intercomparison of GPM and TRMM errors using hourly measurement data on the Tibetan Plateau. Analyzes of the 2014 warm-season three-hour precipitation estimates show that IMERG shows significantly better correlations and smaller errors than 3B42V7. 2011) studied regional and seasonal changes in rainfall systems in Bangladesh using BMD radar data. According to the A1B scenario, towards the end of the 21st century, the frequency and intensity of heavy precipitation events will increase, while the number of wet days will decrease. 2005) describes and compares the large-scale distribution, predominant rain types, and lightning activity of premonsoon and monsoon rain using TRMM-PR and the Lightning Imaging Sensor (LIS) for the years 1998 to 2000 between 36S and 36N. Higher illumination intensity, high echo peak and deep convection were found before monsoon.
They found that high rainfall occurred in the northeastern region (Sylhet) during the pre-monsoon season and the southeastern region (Teknaf) during the monsoon season. Schumacher and Houze (2003) examined the mapping of the shallow, isolated radar ecocategory across the tropics and derive from this mapping a meaningful picture of the overall ensemble of precipitating convection in the tropics. Shallow, isolated rain features dominate the outer fringes of the tropical rain zone, but give way to deeper, more organized convective systems and associated stratiform regions toward the heavy rain areas of the near-equatorial tropics.
The authors found that the version 7 product has significantly less bias and an improved representation of the precipitation distribution. They also found a reduction in the relative bias between observed and simulated fluxes of 30%–95%.
Overview of the study
The wide distribution of droplet sizes usually results from the initial presence of aerosol particles of different sizes. Ice particles grow at the expense of water droplets and then fall, often gaining additional mass by accretion. The central valleys of the Atacama Desert, protected by the surrounding mountains, can go decades without rainfall).
Orographic effects on precipitation are also responsible for some of the planet's sharpest climate transitions. CAPE is the energy available for the free lift of the air parcel from level of free convection (LFC) to level of neutral buoyancy (LNB). CAPE is also a measure of maximum kinetic energy per unit mass of air parcel that can be achieved by moist convection and is therefore an estimate of the maximum possible updraft velocity.
Adams and Souza (2009) showed that the relationship between convective rainfall and CAPE is regionally dependent and mostly not consistent. 1981), for a continental region of eastern Canada, found large positive correlations between areal rainfall and CAPE. The hilly region is found in the southeast of the country and consists of the Chittagong Hills.
Seasons of Bangladesh
During this season, pressure increases over the land from south to north as part of the Siberian High (Skamaruk et al., 2008). In January 2010, the northern and southwestern parts of the country experienced a rapid drop in temperature with cold winds and dense fog. In January 2011, the Meteorological Department recorded the temperature as 2 to 5 degrees Celsius lower than the normal average temperature (about 10°C) (Begum et al. 2015).
Global Precipitation Measuring Mission
Overview of GPM Project
The GPM nuclear satellite was jointly developed by JAXA and NASA, and the DPR was jointly developed by JAXA in collaboration with NICT. The satellite constellation is made possible by an international collaboration for an existing or future microwave imaging station (imager/sonder) aboard a project whose NASA, National Oceanic and Atmospheric Administration (NOAA), European Organization for the Exploitation of Meteorological Satellites (EUMETSAT), National Center D' Etudes Spatiales (CNES) and Indian Space Research Organization (ISRO) etc. Similar to JAXA Earth Environmental's Water Cycle Variation Mission, the GPM mission aims to continue and expand.
The GPM Microwave Imager (GMI)
Only the central parts of the GMI swath will overlap the DPR swaths (and with a duration of approximately 67 seconds between measurements due to the geometry and motion of the spacecraft).
Data Sources and Duration
Data Description
IMERG
The monthly period for 3IMERGM is determined by the typical monthly calendar period of precipitation gauge analyses.
Rain Gauge Data
ECMWF
Methods
- Correlation Coefficient
- Root Mean Square Error (RMSE)
- Standard deviation
- Regional classification
- Classification of different threshold extreme rain
In this study, IMERG data are observed values and rain gauge data are modeled/reference values. In statistics, the standard deviation (SD, also represented by the Greek letter sigma σ or the Latin letter s) is a measure used to quantify the amount of variation or dispersion of a set of data values. A low standard deviation indicates that the data points tend to be close to the mean (also called the expected value) of the set, while a high standard deviation indicates that the data points are spread over a wider range of values.
BMD observation location names over Bangladesh are shown above the station location. Very heavy rainfall (VHR) is defined as the occurrence of rainfall amount greater than 88 mm in 24 hours. The number of very heavy precipitation (VHR) events was calculated using the command =COUNTIF(data range,”>88”).
The number of moderately heavy (MH) precipitation events was calculated using the command =SUMPRODUCT ((data range≥22.5 )*(data range<43.5)). The number of moderate rainfall (MR) events was calculated using the command =SUMPRODUCT ((data range ≥10.5 )*(data range <22.5)).
Variability of rainfall
Daily rainfall
In this figure, the dashed blue line (1:1) is above the solid red line (linear fit), suggesting that the systematic negative change (IMERG < rain gauge) is significant for higher precipitation rates.
Monthly rainfall
The spatial distribution of October 2014 rainfall measured by IMERG and rain gauges is shown in Figure 4.4(e, f). The spatial distribution of IMERG and rain gauge rainfall during April 2015 is shown in Figure 4.5(i, j). The standard deviation of rainfall between IMERG and rain gauge during April 2014 is shown in Figure 4.6(b).
The standard deviation of precipitation between IMERG and the rain gauge in May 2014 is shown in Figure 4.6(d). Precipitation anomalies between IMERG and rain gauge observations in June 2014 are shown in Figure 4.7(a). The standard deviation of precipitation between IMERG and the rain gauge in June 2014 is shown in Figure 4.7(b).
Rainfall anomalies between IMERG and rain gauge during July 2014 are shown in Figure 4.7(c). The standard deviation of rainfall between IMERG and rain gauge during July 2014 is shown in Figure 4.7(d). The monthly rainfall anomaly measured by IMERG and the rain gauge during August 2014 is shown in Figure 4.7(e).
The standard deviation of precipitation between IMERG and rain gauge in August 2014 is shown in Figure 4.7(f). The standard deviation of rainfall between IMERG and rain gauge during September 2014 is shown in Figure 4.7(h). November 2014 precipitation anomaly observed by IMERG and rain gauge is shown in Figure 4.8(c).
The standard deviation of rainfall between the IMERG and rain gauge during November 2014 is shown in figure 4.8 (d). The standard deviation of rainfall between the IMERG and rain gauge during December 2014 is shown in figure 4.8 (f). The standard deviation of rainfall between the IMERG and rain gauge during April 2015 is shown in Figure 4.9(h).
Seasonal rainfall
The trend of monthly variation of precipitation observed by IMERG and rain gauge is almost the same. The maximum rainfall observed by IMERG and rain gauge in Khepupara (1812.09 mm) and Teknaf (3310.6 mm) respectively. Both IMERG and the rain gauge observed minimum rainfall in Chuadanga (200.52 mm by IMERG and 209.63 mm by rain gauge).
The spatial distribution of post-monsoon rainfall measured by IMERG and rain gauge is shown in Figure 4.11(e, f). On the other hand, the lowest rainfall is observed from IMERG and rain gauge at Rajshahi (19.24 mm from IMERG and 5.1 mm from rain gauge). The spatial distribution of winter precipitation measured by IMERG and the rain gauge are shown in Figure 4.11(g, h).
IMERG observed maximum rainfall at Sylhet stations (98.63 mm) and least at Chuadanga stations (8.97 mm). On the other hand, the rain gauge recorded maximum rainfall at Jessore stations (95 mm) and minimum rainfall at Rangamati stations (1.1 mm). The greatest underestimation is in the Sylhet region and the greatest overestimation is in the Rangpur region.
The anomaly of monsoon rainfall observed by IMERG and rain gauge is shown in Figure 4.12(b). The maximum underestimation is found in the Teknaf region and overestimation is found in the Khepupara region. The rainfall anomalies from IMERG and BMD observation in the post-monsoon season are shown in Figure 4.12(c).
From the figure, it is clear that IMERG overestimated in the northern region and underestimated in southern region except some coastal areas. The disparity in rainfall during the winter measured by rain gauge and IMERG is shown in Figure 4.12(d). IMERG is overestimated in the eastern and northeastern parts while it is underestimated in the western part of the country.
Annual rainfall
IMERG is overestimated in the central and northwestern parts of the country while it is underestimated in the rest of the country.
Dry and wet regions
Rainfall events
Threshold rainfall events
Different percentiles of rainfall events
Reason of overestimation in pre-monsoon and underestimation in monsoon
The average daily rainfall measured by IMERG and rain gauge is found to be 4.18mm and 5.61mm during the study period. Comparison of integrated multi-satellite retrievals for GPM (IMERG) monthly precipitation and rain gauge data.
