1. Introduction
Geographic features significantly characterize the spatial pattern of local precipitation (e.g., Houze 2012; Roe 2005). Finescale spatial deviations in precipitation have been investigated by numerous researchers for specific domains according to regional needs (e.g., Isotta et al. 2014). The downscaling of observations and model simulations is required by society (e.g., Duan et al. 2015; Fang et al. 2013; Ménégoz et al. 2013; Park 2013; Qian 2008; Shi and Song 2015). Smoothing algorithms or interpolation techniques can produce high spatial resolutions; however, issues related to the steep climate gradient remain (e.g., Hijmans et al. 2005; Mair and Fares 2011).
Long-term data gathered by a Precipitation Radar (PR) onboard the Tropical Rainfall Measuring Mission (TRMM) are enabling steady progress in explaining finescale precipitation climatology over land and oceans (e.g., Biasutti et al. 2012; Nesbitt and Anders 2009). The spatial resolution of the mean rainfall map is now being shifted to 0.05°–0.1° by maximizing the information in an instantaneous footprint size of approximately 5 km (e.g., Anders and Nesbitt 2015; Bhatt and Nakamura 2006; Biasutti et al. 2012). Compared with products from other satellite sensors, the long-term TRMM PR data have obvious advantages for detecting distinct orographic and coastal rainfall. Furthermore, research on precipitation system climatology has been conducted to understand rainfall regimes embedded in averaged statistics (e.g., Hirose et al. 2009; Houze et al. 2015; Liu et al. 2008; Nesbitt and Anders 2009; Rasmussen et al. 2016; Roca et al. 2014). These previous studies described the regional variation of gridded rainfall as an ensemble of precipitation systems generally characterized by rainy areas. However, it is not yet completely understood how precipitation system climatology can be resolved at a resolution of 0.05°–0.1°. The reallocation of spatial storm information to finescale gridpoint data is expected to distinguish the geographical constraints on rainfall. Understanding the core characteristics of high-resolution rainfall climatology will help distinguish signals from noises in other datasets, such as the short-term average of the Dual-Frequency Precipitation Radar (DPR) onboard the Global Precipitation Measurement (GPM; Hou et al. 2014) core observatory.
To evaluate the spatial coherency or regional heterogeneity of the average rainfall, additional research is needed to better understand the geographical association of rainfall in terms of its spatial and temporal discontinuities over long periods. Orographic lifting and thermal circulation, including mountain/valley breezes and sea/land breezes, exert a significant impact on the incidence of localized precipitation systems. For example, a number of studies have investigated the diurnal characteristics of rainfall over the northeastern region of Brazil with regard to inland propagation induced by sea breezes (e.g., Angelis et al. 2004; Janowiak et al. 2005) and the river breeze effect (e.g., Silva Dias et al. 2004; Negri et al. 2002). Such phenomena, caused by local circulation, can be resolved only by extensive and homogeneous precipitation datasets with high resolution (e.g., Fitzjarrald et al. 2008). Understanding how characteristic storm groups vary by region and time requires additional evaluation using updated finescale TRMM PR climatology.
For localized storms, rainfall enhancement and lightning concentration over small islands have been previously investigated with regards to dynamics, thermodynamics, and aerosol impacts. Nullet and McGranaghan (1988) investigated the ratio of island rainfall to oceanic rainfall over the Hawaiian Islands and noted uncertainties in regional datasets based on the gauge network. Yang and Chen (2008) simulated the effect of terrain height and island size on regional rainfall in Hawaii and showed a complex, orography-related effect. Of the global precipitation datasets for coastal or island rainfall, the long-term mean TRMM PR data are the most useful. Heiblum et al. (2011) discussed precipitation formation in the coastal regions of the eastern Mediterranean. Ogino et al. (2016) highlighted the significance of locally concentrated precipitation for entire tropical coastlines. Sobel et al. (2011) conducted a study on the scale and topographic dependencies of rainfall enhancement over small islands with areas between 100 km2 and several thousands of square kilometers in both the Indo-Pacific Maritime Continent and Caribbean regions. They found that rainfall over small islands with areas larger than 315 km2 exhibit land-type characteristics. Williams et al. (2004) proposed that low-relief islands smaller than 100 km2 exhibited oceanic behavior. Conversely, moisture convergence and the resulting rainfall enhancement are expected to some extent even over very small islands, as indicated by the island size sensitivity experiments conducted by Cronin et al. (2015). Storm characteristics over very small islands in the global tropics have received less attention. An overall analysis of the spatial representativeness of rainfall over low islands of coral origin is still necessary to demonstrate offshore rainfall (e.g., Wang et al. 2014). A question also remains concerning the significance of the land–sea contrast of rainfall around small volcanic islands with topographic obstacles.
As mentioned above, several studies have been made of the spatially coherent characteristics based on 0.05°–0.1° gridded datasets. However, little is known about the steepness of the spatial gradient, the importance of higher-resolution data, the advantage compared to the coarse statistics (e.g., at 0.5° scales), and local-specific retrieval errors. The lack of an in-depth understanding of kilometer-scale features poses an obstacle to multiscale climatology and advanced observation strategies.
By compiling a large number of high-impact storm snapshots, Hirose et al. (2017) demonstrated that 16 years of TRMM PR data at a 0.1° scale resolved coherent geographic rainfall patterns. In our study, this dataset is used to examine precipitation system climatology at fine scales, with a particular focus on the discontinuous variation around high mountains, the Amazon River, very small islands, and coastlines. The presented study examines the geographical features of prevailing precipitation systems in terms of local mechanisms related to the underlying topography or geography, with a special focus on the 0.1°-inherent spatial fluctuation. This paper also investigates spatially fixed retrieval uncertainties.
2. Data and methodology
a. Retrieval differences of rainfall over land and oceans
This study primarily uses estimated surface rainfall data derived from the TRMM PR 2A25 version 7 algorithm (Iguchi et al. 2009) for 1998–2013. Even though PR uniformly observes radar reflectivity over land and water (e.g., McCollum and Ferraro 2005), there are several factors that can potentially affect the land–ocean contrast of the estimated surface rainfall. One such factor could occur during the path-integrated attenuation estimation (Seto and Iguchi 2007; Takahashi and Iguchi 2007). Moreover, differences in storm statistics and surface conditions have resulted in different incidence-angle dependencies (Hirose et al. 2012). An angle-bin difference of several percent remains in the PR 2A25 version 7 rainfall data, whereas the overland difference has decreased by nearly half compared with the version 6 data (see the JAXA TRMM PR V7 validation website; http://www.eorc.jaxa.jp/TRMM/documents/PR_algorithm_product_information/pr_v7_validation_jpn_e.htm). The geolocation data archived in the standard product are assumed to be valid when analyzing features finer than the footprint size (Takahashi and Oki 2010). The geometric adjustment for data at the surface is negligible over low-lying land and oceans but affects the statistics slightly over alpine terrains. For example, the largest horizontal difference in a location at 6000 m in altitude between the actual surface and the geolocation data projected onto the Earth rotating ellipsoid is approximately 1.8 km, less than half the footprint size.
Given the wide range of climatological uses, the remaining unsolved issues are crucial factors to be incorporated into global water studies to obtain the current best estimates. This study explores not only how finescale TRMM PR data are superior to other data but also what kind of retrieval uncertainties exist in representations of geographic features at the 0.1° scale.
b. A database of PR-captured precipitation systems
Instantaneous rainfall information was compiled into a high-resolution gridded dataset sorted by scale-based precipitation systems (Hirose et al. 2017). The number of gridded samples was taken into account in the calculation of the per-grid rainfall amount. Here, a storm is defined as a spatially continuous rainfall area (Hirose et al. 2009), which is called a PR-captured precipitation system (PR-PS). The PR-PSs are simply categorized as small, medium, or large based on horizontal scales of <10, 10–100, and >100 km, respectively. Localized small PR-PSs have not been well represented in other datasets. Hirose et al. (2017) showed that hundreds of large PR-PSs that significantly characterized the total rainfall pattern were observed over major rainfall areas during the 16 years of TRMM PR observation. According to the geographic information, this study focuses on the relevancy of storms with various spatial scales.
The storm database at a resolution of 0.1° filters out clutter contamination using a removal filter developed by Hamada and Takayabu (2014, hereafter HT14). Hirose et al. (2017) reported that the filter effect is not significant over most areas even though more than 10% of the estimated surface rainfall is sporadically reduced over specific mountainous areas, such as the slopes of the Himalayas. An example of the correction effect on specific orographic rainfall is provided in section 3. This clutter-removed, finescale storm database, together with various rainfall distribution maps, is accessible on our website (http://www.rain-clim.com). Using the zoom-out function and the transparency setting in the options, the finescale supplemental information can be accessed.
c. Islands determined by 0.1° land pixels
One of our primary focuses concerns the discrete distribution of rainfall over land and oceans. Section 3d highlights rainfall over small islands. The target is 0.1°–1°-scale islands in the global tropics, and their scale-based storm statistics are shown. Here, the definition of an island and the dataset attributes are explained. Islands are identified, based on SRTM30 (Farr et al. 2007), as a conglomerate of land pixels for which 30-s pixels with an altitude of >0 m prevail in each 0.1°-scale grid. Averaged and maximum 30-s altitude data are archived for each land grid. In this paper, islands with area-equivalent diameters of 0.1°–1° and 0.1° are termed “small islands” and “very small islands,” respectively. Figure 1 shows a map of the islands and the area categories in the surrounding oceans. Each surrounding ocean is determined by the water area within a 1° distance from each island’s outer grid. The island area is categorized according to the number of 0.1° land pixels corresponding to the island area. They are classified according to the following scale: 0.1° [category 1 (C1)]; 0.1°–0.2° (C2); 0.2°–0.4° (C3); 0.4°–0.7° (C4); and 0.7°–1° (C5). Within 36.5° latitude of the equator, there are 462 small islands. Of these islands, 63% are located in the western Pacific (90°–180°E). Larger islands generally have higher mountains. Of the small islands, 61% have sizes less than or equal to 0.2°. More than half of the 0.1°–0.2° islands have a mean elevation of <100 m, and 24% reach a maximum elevation of <100 m. A total of 94 islands have a maximum elevation of 1–99 m, 187 have a maximum elevation of 100–499 m, 114 have a maximum elevation of 500–999 m, and 67 have a maximum elevation greater than or equal to 1000 m. The number of very small islands in each maximum elevation range is 54, 100, 27, and 3, respectively.
The geographical distribution of the islands. Color shading for each adjacent ocean indicates the island area in terms of the number of 0.1° pixels.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
The pale purple shadings in Fig. 1 also show the distribution of 15 islands larger than 1° but smaller than the Australian Continent. Section 3e examines the coastal rainfall surrounding the aforementioned small islands, other large islands, and continents. The maximum elevations of three-quarters of the large islands exceed 1000 m. Over the TRMM domain, New Guinea is the largest island that includes the highest elevation. The rest of the 0.1° land pixels belong to continents whose coastlines are indicated by the thin black lines in Fig. 1.
3. Finescale precipitation system climatology
This section describes the local dependency of precipitation climatology at the 0.1° scale, the diurnal variations, and the constituent behavior of scale-based precipitation systems.
a. Localized variations in the rainfall
Figures 2a–2d indicate that the identification of underlying small spatial-scale features is sensitive to the data resolution. Figure 2e represents the percent slope based on a neighborhood algorithm using 0.1°-altitude data averaged from SRTM30. These long-term high-resolution data allow us to update our views with regards to orographic rainfall. Scale drawings can be found on our website, which was introduced in section 2. Two narrow high-rainfall bands (Shrestha et al. 2012) are identified along the Himalayas only in Fig. 2d. Moreover, Fig. 2d delineates the extent of the heavy rainfall regions by southwesterly moist airflows that impinge upon the Western Ghats, the western region of the Indian subcontinent. Xie et al. (2006) reported that the rainfall maximum is tightly trapped along the coast. The long-term finescale TRMM PR data confirm the locally fixed features in a more refined sense. As demonstrated in other studies, such as Romatschke and Houze (2011) and Biasutti et al. (2012), rainfall concentration around the western coast of the Indian subcontinent appears as a very narrow rainband over the windward slope of the Western Ghats. Rainfall of more than 8 mm day−1 is captured, along with a narrow band where the slope ranges from 1% to 4% (with an average slope of 2.2% and altitude of 369 m). Areas with rain greater than 5 mm day−1 accounted for only 15% of the southwestern area of India (12°–18°N, 72°–78°E) but accounted for 45% of the rainfall amount. Note that the amount of rainfall over the coastal lowland was approximately half that generated on the slopes. This is attributable to local differences in orographically forced vertical motion (Shige et al. 2014).
Rainfall distribution over Asia at (a)–(d) different resolutions, with enlarged boxes showing rainfall on the west coast of India and Myanmar. (e) The spatial gradient of the altitude based on the neighboring eight 0.1° elevation data points.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
By contrast, for areas around the Arakan Mountains and the Tenasserim range, rainfall maxima appear over the nearby ocean. Over the past few decades, the generation of conspicuous offshore convection has been investigated in terms of large-scale environments and mesoscale circulation (e.g., Grossman and Garcia 1990; Hirose and Nakamura 2005; Houze et al. 2015). The exact location of maximum rainfall is still controversial (Biasutti et al. 2012). This dataset shows that the peak annual rainfall is approximately 14 mm day−1 and is observed over the coastal ocean of Myanmar (16.4–16.5°N, 97.3–97.4°E). A high-rainfall area (>10 mm day−1) is widely distributed over the coastal ocean. In addition, a local rainfall maximum of 13 mm day−1 is observed at a narrow inland area near the Dawna range in eastern Myanmar. Over the eastern part of the Gulf of Thailand, a widespread rainy area is observed. However, on a finescale, peak rainfall appears over coastal landmasses to the windward side of the Cardamom Mountains in the southwest of Cambodia and eastern Thailand. Finescale data can shed light on these sharp geographic dependencies by region.
The following examples illustrate rainfall in alpine environments. One of the most plentiful rainy regions with a rugged terrain is found on the island of New Guinea. A localized low-rainfall region is located over the summit of the Western Sudirman range, in the vicinity of the first and second highest mountains in Oceania. Such low precipitation areas around mountain summits have been detected over tens of specific mountains; for example, they are found around the three highest peaks in Africa (Kilimanjaro, Mt. Kenya, and Rwenzori; all >5000 m; figure not shown) and high mountain peaks over the Himalayas. Specific data features around the Himalayas will be described in the next section.
Some regions of Asia and Central America with steep terrain have high rainfall fluctuations. To highlight the distinct climatic variations across the divide, Fig. 3 illustrates the spatial nonuniformity of rainfall, as defined by a significant increase or decrease in the 0.1°-scale rainfall over its neighboring pixels. This figure clearly indicates the local rainfall concentration and adjacent semiarid areas over two high rainfall-varying regions: the northwestern part of South America and Southeast Asia. The cool-colored (warm-colored) shadings indicate highly concentrated (locally low) rainfall areas where rainfall within the grid is at least 3 mm day−1 greater (less) than that over more than 25% of the adjacent areas within 0.5°. Rain shadow areas, which generally occur on the leeward side of mountain barriers, are identified by yellow shading, in combination with blue shading in the neighboring high-rainfall areas. They appear on the west coast of the Indochina Peninsula, where abundant rainfall is observed offshore (Fig. 3b). Conversely, yellow shading appears on the windward side of the coastal oceans of Colombia (Fig. 3a) and the foothills of the central mountain range on the island of New Guinea (Fig. 3b), which exhibits an abrupt increase in rainfall. High- and low-rainfall areas are localized in coastal landmasses and mountainous terrain, except for a small fraction of coastal waters around the Philippine Archipelago and Sumatra and the closed-off sections of several bays in the tropics. The high spatial discontinuity and spatially fixed patterns are important for understanding the mechanisms of orographic precipitation and for assessing the need for dense and accurate observations of local water budgets and practical uses. The island of New Guinea and the northern Andes, identified as areas with the sharpest spatial gradients of rainfall, could be appropriate sites to verify finescale precipitation data. Therefore, high-resolution precipitation data reveal the need for various types of high-resolution information.
Fraction of surrounding pixels with differential rainfall ≥ 3 mm day−1 over two distinct areas: (a) the northwestern part of South America and (b) Southeast Asia. Rainfall over the blue (red) color is significantly high (low) relative to that over the surrounding pixels within 0.5°.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
Maximum annual rainfall at a 0.1° scale is observed at the specific regions listed in Table 1. Total rainfall has its maximum in the western region of the Alejandro Selkirk Island of Chile. Judging by the extremely high concentration at this specific point, these signals are likely deteriorated by ground clutter that was not sufficiently removed by the HT14 filter. The second highest value (25 mm day−1) is observed over the ocean near the curved coast of the western region of Colombia, close to the location where large PR-PS rainfall exhibits a maximum. Maximum rainfall according to medium PR-PSs is observed on the southern coast of Bioko Island, Equatorial Guinea, west of Cameroon. Therefore, the extremes of the large and medium PR-PSs are likely associated with orographically modified flow (e.g., Biasutti et al. 2012; Mapes et al. 2003). By contrast, the extremes of the small PR-PSs are more difficult to explain. They appear as enormous isolated peaks over steep orographic islands in low-rainfall oceans or high-altitude areas. This can be seen in the finescale rain climatology over the Alejandro Selkirk Island, Chile, where PR-based 0.1° rainfall is approximately 50 mm day−1 (even after filtering, and nearly the same value is seen in the version 6 data). The HT14 filter removed only 10% of the rainfall in this area. This extreme value over the low-rainfall ocean results in the current 0.1° rainfall maximum over the TRMM domain. When a PR-PS is composed of more than one footprint rain sample (not shown) 30% of the abundant value over the Alejandro Selkirk Island can be reduced. However, a minimum threshold storm size is not adopted in this paper because the correction performance is insufficient, and it drastically deteriorates the detection of small PR-PSs in other locations. This issue, found in the finescale climatology, is described more in the following section.
Location of the highest annual rainfall from scale-based systems. The spatial resolution is 0.1°. The areas of suspicious statistics with enormous isolated rainfall peaks are enclosed in parentheses.
b. Precipitation systems over the Himalayas
The local characteristics of precipitation systems around the peaks of the Himalaya range in central Asia were investigated to demonstrate the discrete pattern of the significant barrier effect over areas where ground- or space-based estimates have clear uncertainty. In this section, precipitation in summer [June–August (JJA)] is examined to reduce the uncertainty in the snow detectability due to TRMM PR’s low sensitivity. Figure 4a highlights the rain fraction of small PR-PSs. The clutter filter removes some isolated echoes, primarily at an altitude of 1000 m, but does not affect the spatial pattern. The rain fraction arising from small PR-PSs is approximately 1%–2% over the southern foothills and monsoon trough regions, whereas it increases to 10%–20% over the Tibetan Plateau because of the higher and lower occurrence frequencies of small and large PR-PSs, respectively. The highest concentration (>90%) appears around the highest mountain ranges at the northern border between Nepal and China. This pronounced concentration is found over mountains higher than 4000 m. In particular, 95% of the rainfall is from small PR-PSs over a domain near Mt. Everest (27.95°N, 86.85°E). To the south of Everest, rainfall from small PR-PSs is estimated to be approximately 400 mm from June to August.
Finescale rain features around the highest range of the Himalayas in JJA: (a) the rain fraction of small PR-PSs, (b) the number of small PR-PS centroids for each 0.01° grid, (c) differential rainfall between inner-swath TMI (>0.05 mm h−1) and PR rainfall, and (d) topography. The black line indicates the border. The red line in (d) indicates 4000 m in altitude.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
Figure 4b exhibits distinct small-storm concentrations at the centroid location of individual small PR-PSs. The circularly arranged rainfall pattern does not appear as a barrier-scale effect on the windward slope, except along valleys and in the neighborhood of high mountain peaks, such as Everest, Lhotse, and Makalu. The averaged altitude of this small PR-PS concentration is approximately 3900 m, even though the altitude has an irregular contour that is distinct from the circle (Fig. 4d). Similar storm-concentrated circles appear over particular individual mountains, such as Mt. Fuji in Japan. Yasunari and Inoue (1978) reported that a precipitation maximum was observed from approximately 4000 m in altitude to the terminus of a glacier (5200 m). In addition, Harper and Humphrey (2003) used the glacial ice flux model to suggest that precipitation reaches a maximum level well below the elevation of the highest peaks. Our results for micro rain climatology are mostly consistent with such results in that the upper 1000+ m of the massif is a high-elevation desert with little precipitation. Conversely, even though this is a summer case, retrieval uncertainties, particularly for snowfall around Mt. Everest, could contribute to the high-altitude low precipitation because of limitations in the sensor sensitivity, the missing effect from the clutter mask adjacent to the surface, and possible discrepancies in the hydrometeor assumption. Most of the ring radii are slightly larger than the footprint size. The number of ring patterns has a moderate annual cycle with a summer peak. However, the altitudes vary little with season. The ring patterns likely suffer from interference from adjacent areas that are outside the effective field of view. These multiple uncertainties over complex orography should be further evaluated to explain alpine climatology.
The difference in rainfall between the TRMM Microwave Imager (TMI) 2A12 product and the PR datasets is examined here. To compare with the PR rainfall, the minimum threshold of the TMI rainfall was set to 0.05 mm h−1 (Seo et al. 2015). This extremely weak rainfall, which was less than the threshold, had a major impact on oceanic rainfall but a very small effect over land. The TMI rainfall data were constructed on the basis of the PR-swath range. Figure 4c indicates that significant retrieval differences remain to be mitigated, particularly for orographic rainfall where the spatial gradient of the rainfall is sharp. The TMI rainfall over the Himalayas and the foothill regions is overestimated and underestimated, respectively, when compared with the PR rainfall. In winter, the TMI rainfall is underestimated over alpine regions (not shown). These differences are partially attributable to uncertainties at the lowest end of the Gaussian estimate, snow-cover contamination in the TMI estimates, differences in the field of view, PR main-lobe clutter contamination, and near-surface profile model mismatches both in PR and TMI (e.g., Duan et al. 2015; Liu and Zipser 2014; Zagrodnik and Jiang 2013; Shige et al. 2013; Yamamoto et al. 2017). The above-mentioned small PR-PS rainfall, which is high over the Himalayan transitional zone and alpines (e.g., to the south of Everest), is occasionally greater than the TMI rainfall despite the tendency for TMI overestimation in this region. TRMM PR rainfall climatology and its in-depth verification could be a benchmark to detect spatially extended orographic rainfall and detailed local uncertainties; this would be valuable for mutual algorithm development based on active and passive sensors and satellite and ground observations.
The localized rainfall around Everest is hard to detect using other satellite or rain gauge data (e.g., Andermann et al. 2011; Dhar and Nandargi 2000; Immerzeel et al. 2015; Yamamoto et al. 2011). High-resolution PR rain climatology can be used as a detector of signals in this region. The precipitation rings around the summits might result from orographically and thermally induced cumulus convection (e.g., Ageta 1976; Tartari et al. 1998). However, before entering into a detailed discussion, note that such locally trapped PR echoes might be affected by surface clutter, as implied by the low-level echo profiles studied by Takahashi and Oki (2010) or by a mismatch in the digital elevation model data (HT14). For this dataset, a conservative clutter filtering was adopted; however, the moderate clutter contamination potentially remains. From a quantitative consideration of the snowfall over the peaks, the use of other highly sensitive sensors, such as DPR onboard GPM, is expected to evaluate the TRMM PR’s detectability of snow profiles in the future. Moreover, the localized PR rainfall should be further evaluated in view of locally fixed retrieval errors, including missing shallow storms and mismatches in the assumed downward-decreasing profile model under thick main-lobe clutter contamination (Hirose et al. 2012).
c. Suppression of rainfall along the Amazon River
Figure 5 illustrates local rain patterns over certain parts of Latin America. In contrast to the conspicuous spatially fixed orographic effect around Central America and near the Andes (Biasutti et al. 2012; Rasmussen et al. 2016), the spatial gradient of annual rainfall around the Amazon River basin is moderate (Fig. 5a). Seen with topography, as depicted in Fig. 5c, the annual rainfall is modulated even for gently sloping hills with altitudes of 200 m. Moreover, the finescale climatology leads to a slight reduction in the rainfall along upstream areas of the river. The contrasting characteristics around inland waters differ from the local concentration (such as on Lake Maracaibo and Lake Victoria). Rainfall suppression along the meandering stream becomes sharper in the afternoon [1200–1800 local time (LT)], as shown in Fig. 5b. In particular, the lowest afternoon rainfall in the Amazon basin occurs near Santarém, which is located at the intersection of the Amazon and the Tapajós, in conjunction with the river breeze and inhibition of the sea-breeze front parallel to the coastline (e.g., Fitzjarrald et al. 2008; de Oliveira and Fitzjarrald 1993). In contrast to the concentration of >40% (60%) for the surrounding regions (the Atlantic coastal land), the rainfall fraction at 1200–1800 LT over stream waters with widths of a few tens of kilometers is less than 10% (not shown).
Rainfall over the Amazon basin: (a) annual and (b) afternoon rainfall (1200–1800 LT). (c) The SRTM30 elevation. The box inside (c) indicates the selected region discussed in Fig. 6.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
With regard to temporal variations in the rainfall, a narrow band is discernable from the diurnal dependence at the midstream of the Amazon River, in particular for medium PR-PSs of 10–100 km (Fig. 6a). Here, pixels with negative (positive) anomalies in the 1-h period before or after a period of maximum (minimum) rainfall were excluded as temporally incoherent peaks with sharp fluctuations due to insufficient sampling (Hirose et al. 2009). The time of maximum (minimum) rainfall over water is associated to some extent with the time of minimum (maximum) rainfall over adjacent areas, as is the case of the coastal area around the mouth of the Amazon River. The suppression of afternoon rainfall by small PR-PSs is detected, albeit slightly, along the river. The suppression of afternoon rainfall by medium PR-PSs occurs earlier in the middle of the river than at the edge. The localized medium PR-PSs characterize the diurnal variation of rainfall around the river, whereas the westward-migrating large PR-PSs have a significant impact on the diurnal features over other areas in this region.
Diurnal variation in the rainfall due to scale-based PR-PSs around the midstream of the Amazon River. (a) Time of maximum and minimum hourly rainfall. Temporally incoherent peaks are excluded, as noted in the text. (b) The latitude–time cross section of rainfall at 55°W.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
Figure 6b shows the discrete diurnal pattern across the river along 55°W for the three precipitation system size categories. The rainfall maximum appears at approximately 0600 LT close to the river (near 2.5°S). Small PR-PSs generally develop over land in the early afternoon, and the daytime peak disappears over the river. Instead, a slight enhancement is detected in the early morning. Rainfall due to medium PR-PSs fluctuates with a pronounced seesaw pattern close to the river. Afternoon and morning rainfall due to medium PR-PSs dominate over the land and river, respectively. By contrast, the latitudinal river effect is unclear for large PR-PSs. Large PR-PSs bring morning rainfall to this area (e.g., Kousky 1980). The all-time averages show few features on the river because of the combination of suppression and enhancement. Dividing the data into hourly bins and storm spectra enhances the detection of localized features on the river.
Signs of the decrease over the Amazon River were clearer in the TMI data than in the PR data, even though this is not shown here. Retrieval uncertainties in the TMI version 7 (V.7) rainfall remain, particularly over mixed surface conditions. Because of the background surface, the different procedures make it difficult to extract the finescale variation around the water (McCollum and Ferraro 2005; Mega and Shige 2016; Tian and Peters-Lidard 2007). Mitigation is of great importance in detecting dynamic features. Further investigation into the retrieval and sampling errors of various products is required to discern local characteristics at higher spatial and temporal scales.
d. Rainfall anomalies over very small islands
This section highlights the detectability of finescale rainfall features over very small islands. The definitions of land pixels, islands, and the surrounding ocean are described in section 2c. Figure 7 and Table 2 show the spatial anomalies of rainfall over islands compared with the surrounding ocean [here called the rainfall enhancement (RE)]. The RE is the difference in the rainfall between the island Risland and the surrounding ocean Rocean divided by Rocean. On average, for C1 and C5 islands, rainfall is increased by approximately 5% and 25%, respectively. The RE for small PR-PSs exceeds 50% over 44% of the very small islands. In particular, nine-tenths of very small islands with a maximum elevation > 500 m have an RE of more than 50%. The RE is increased by more than 80% over very small islands with high mountains (>1000 m). Even with flat (all) C1 islands, a 26% (71%) positive anomaly is detected in the RE as a result of small PR-PSs. The REs for larger islands (C2–C5) are significant for medium PR-PSs, reflecting the development of convection over land. By contrast, the spatial anomalies resulting from large PR-PSs (except for small islands with high mountains) are negative. For small tropical islands, locally enhanced precipitation systems are likely to be smaller than the island size. These differences will be further discussed in terms of the diurnal cycle in the next section.
Spatial anomalies of the rainfall over scale-based islands as compared with the surrounding ocean. Anomalies based on averaged rainfall for each category (C1–C5) are illustrated by lines and dots for each island. Thick gray lines indicate the statistics regardless of elevation. Colors show the maximum elevation of each island.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
RE (%) over scale-based islands. The first line indicates the island properties on the maximum elevation.
These results are based on the HT14-filtered product. A comparison with results based on the unfiltered product has been made (not shown). The filtering impact is nearly zero for flat islands, even though some effect is detected over volcanic islands. For example, the RE over very small flat islands is 1% for all systems and 26% for small systems, regardless of filtering. Over very small islands with high mountains (>1000 m), this RE value was 84% (125%) for all systems and 461% (578%) for small systems after (before) filtering. The remaining artificial echoes in the 2A25 version 7 product should be mitigated to allow for the generation of comprehensive and accurate maps of the local climatology. Nevertheless, the results presented here are assumed to be robust at least with regards to bugs in the version 7 clutter routine found by HT14, because the tendency is nearly the same as that based on the version 6 product.
Looking at individual islands, an RE of more than 800% is observed for small PR-PSs over four small islands with volcanoes higher than 1000 m in the Canary Islands, and Cabo Verde, off the west coast of the Sahara. The RE factor over the Fogo Island of Cabo Verde is 27. The localized surface rainfall rate is fairly strong for small PR-PSs. The storm-top height of rainfall concentrated on the windward side is higher than the maximum elevation in this region. The vertical structure of small PR-PSs is unique and has a profile with a steep downward-increasing rainfall rate in the lower atmosphere. There was no significant bias due to the month or local hour (not shown). These features may imply that the dynamic effect is the prevailing mechanism; however, they could also be due to clutter contamination that could not be removed by the HT14 filter. This extreme record based on the current 0.1°-scale dataset needs to be confirmed in future studies.
e. Diurnally varying precipitation systems across the coastline
Rainfall has high fluctuations in time and space around coasts according to the geographical environment (e.g., Bergemann et al. 2015; Mori et al. 2011; Qian et al. 2012). To understand the land–ocean contrast discussed in the previous section, the dynamic state of the coastal rainfall is now examined. This paragraph examines the latitudinal difference in the cross-shore variation to distinguish the various statuses of the local circulations. In Fig. 8, the spatial variation in the hourly rainfall is designated for each distance from the coastlines at four different latitudes. Here, the coast is defined as a 0.1° pixel that contains both ocean and land information (based on SRTM30). A significant diurnal amplitude appears in the low latitudes due to the strong solar insolation. Close observation shows that the relatively narrow transition zone between land and ocean appears only in the midlatitudes between 30° and 36°. Conversely, at latitudes less than 30°, widespread amplifications of the rainfall across the coastline are seen as indicators of the significant inland penetration of the sea-breeze fronts. Therefore, the land–ocean contrast depends on whether the latitudes are higher than 30°. The discontinuity appearing in the diurnal signatures is consistent with a theoretical study of sea-breeze circulations by Rotunno (1983). Over the coastal land, the second to fourth panels of Fig. 8 have common features of the rainfall migration originating from the coastline from 1200–1300 LT to 1900–2000 LT with its maximum at 1500–1600 LT. In other words, the large-scale circulations appear to be nearly in phase with the diurnal heating cycle in the tropics and subtropics, contrary to the simple model predictions (Yan and Anthes 1987).
Hourly rainfall variation around the coasts for latitudes between (top)–(bottom) 30° and 36°, 20° and 30°, 10° and 20°, and 0° and 10°. A positive x-axis value indicates the landward side of the coast. The thick black lines indicate the daily mean.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
Subsequently, coastal rainfall variations according to land area and storm scale have been investigated for latitudes less than 30°. The right-hand panels of Fig. 9 show the hourly rainfall variation around the coasts of the continents (coast C) in the tropics and subtropics (30°S–30°N). The gross features of the peak time and its migration from the coastline are similar to those of the distance from the nearest coast, as shown by Hirose (2015). The top panel of coast C shows that rainfall over land rapidly increased in the early afternoon, and the maximum hourly rainfall occurred 0.3° inland from the coast at 1500–1600 LT. The very narrowly concentrated short-term rainfall is approximately double that of the averaged peak near the coast. Over the coastal ocean, rainfall begins to increase at 2000 LT and reaches its maximum 0.3° offshore at 0400–0500 LT. On average, rainfall is greater over the coastal ocean than over coastal land. A seesaw pattern of land and ocean rainfall is found for coastal regions within approximately 1° of the coastline. Diurnal patterns 3° from the coastline are similar to those for the open ocean or inland areas.
Hourly rainfall variation around the coasts of very small islands with a scale of 0.1° (coast S), small islands with scales of 0.1°–1° (coast M), islands greater than 1° (coast L), and continents (coast C) for latitudes between 30°S and 30°N. A positive x-axis value indicates the landward side of the coast. The rainfall suffixes, A, S, M, and L, indicate all, small, medium, and large systems, respectively. The thick black lines indicate the daily mean.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
The second through fourth panels of coast C in Fig. 9 illustrate the rainfall statistics for scale-based systems. Rainfall from small PR-PSs gradually decreases over hundreds of kilometers, from offshore to inland, with a discrete change in areas within 0.2° from the coast. The mean rainfall slightly increases on the coastline as a result of the significant increase in the afternoon rain over land because the grid designating the coast includes both land and ocean pixels. Small PR-PS rainfall is primarily observed for a few hours in the early afternoon over land. The time of minimum rainfall over the coastal ocean within a distance of approximately 2° from the coastline is different from that further offshore. Therefore, the enhancement in small-scale systems is significantly affected by geographic patterns. Medium-system rainfall noticeably increases 0.2° inland from the coast and characterizes the peak of the total hourly rainfall. Averaged rainfall is spatially coherent offshore and deep inland; however, the diurnal amplitudes are strongly modulated by the land–ocean contrast. The spatial gradient of the hourly rainfall indicates that precipitation formation is particularly favored in areas within 1° of the coastline. Conversely, total hourly rainfall due to large PR-PSs undergoes discontinuous variation at the coast. Diurnal variation over the coastal ocean is significant, with its peak migrating from the coast to the open ocean after sunset. The rainfall has its maximum in the early hours of the morning, which results in the prominent offshore rainfall concentration. The rainfall peaks migrating inland become obscure at midnight. In areas 3° from the coast, variation due to the land–ocean interaction is insignificant for the average rainfall but is detectable for the diurnal amplitude pattern. The diurnal march over several hundreds of kilometers may be modulated by the significant offshore extent of the land breeze in the tropics (Gille et al. 2005).
A statistical assessment of the small island effect was performed in terms of the cross-shore variation and the diurnal cycle. The left-hand panels of Fig. 9 show the case for very small islands (coast S). The island size is obtained from the continuous land pixels. In this figure, the coastal grid, including both land and ocean pixels, is regarded as the very small islands. Only small PR-PSs can provide land-only rainfall for coast S. As for the time of the maximum small PR-PS rainfall, moderate daytime peaks can be detected. Small and medium precipitation systems have a positive impact on rainfall across very small islands; however, the opposite is true for large systems. Rainfall over the adjacent ocean is not significantly modulated. The results for 0.1°–1.0°-scale islands (coast M; approximately the size of medium PR-PSs) are shown in the second panels from the left in Fig. 9. These panels underline the features seen for very small islands. The average rainfall due to small PR-PSs has its peak around the coastline, similar to other land. Over coastal land, the rapid increase in evening rainfall from medium-scale systems is outstanding, and a distinct, negative anomaly from large systems contributes to a lower RE. It is interesting that large PR-PS rainfall is significantly affected by the local environment at fine scales. The morning rainfall concentration over the coastal ocean only appears for waters close to a landmass (>1°), indicating the possibility of a weak moisture convergence resulting from moderate land breezes. An evening suppression over coastal waters is not observed, indicating the likelihood of weak compensating subsidence in this region. Therefore, the RE values over small islands are generated by abundant afternoon rainfall due to medium PR-PSs and the moderate development of large PR-PSs over the coastal ocean.
Conversely, the land–ocean contrast for large islands (coast L, the third column from left in Fig. 9) is different from that for small islands or continents. Rainfall monotonically increases from offshore to inland, except for a slight decrease on the coastline. The landward increase responds to the remarkable enhancement in the morning rainfall over the coastal ocean, the significant afternoon rainfall over the coastal land, and the intensification in the nighttime rainfall migrating inland. Rainfall over land for coast L is more than double that for coast C, which reflects the combined effect of coastal curvature and mountainous topography. The landward increasing trend counteracts the decreasing trend of small PR-PS rainfall shown in coast C. The spatial pattern of medium PR-PS rainfall for coast L is intermediate between that for small islands and continents; however, the increase in the nighttime rainfall is characteristically observed over the island interiors. Over land, the time of maximum hourly rainfall due to large PR-PSs agrees with that for coast C: 1600–1700 (2300–2400) LT at 0.3° (1.3°) inland from the coastline. When compared with coast C, large PR-PSs provide greater rainfall amounts in particular near midnight over inland and in the early morning over the coastal ocean.
The diurnal behaviors in the midlatitudes differ substantially from those in the tropics. Figure 10 illustrates the spatial and diurnal features of coastal rainfall in the midlatitudes. There is no coast-induced rainfall for coast C, except for a slight increase on the coast. The transition between land and ocean appears within 0.3° from the coastline. The effect of local breeze circulation is likely trapped near the coast in the midlatitudes (Rotunno 1983). The rainfall for coast S has no peak over islands. For coast M, small and medium PR-PS rainfall has its maximum over islands as it does in the tropics. However, the negative anomaly due to large PR-PSs outweighs this effect. Coast L in the midlatitudes between 30° and 36° corresponds to the coast of the Japanese archipelago (Fig. 1). Afternoon rainfall over the inland is observed for small and medium PR-PSs. However, the averaged rainfall shows a gentle curve with its maximum over offshore regions, which are determined by large PR-PSs as a result of the high fraction of widespread systems.
As in Fig. 9, but for latitudes between 30° and 36°.
Citation: Journal of Climate 30, 11; 10.1175/JCLI-D-16-0442.1
4. Discussion and conclusions
In this study, our primary focus was to demonstrate the local characteristics of high-resolution precipitation climatology and to understand the adequacy and uncertainty of global rainfall estimates. The results of spatial rain features that depended on terrain forcing were, to some extent, consistent with previous studies. Local representation was refined through the dynamic state of integrated high-impact systems responsible for most of the regional rainfall characteristics (Hirose et al. 2017). This study examined finescale rain concentration in conjunction with orography and geography.
Finescale PR data permitted a discussion of the accuracy of precipitation over unexplored areas, such as alpine regions. Even though Palazzi et al. (2013) illustrated that the gross precipitation climatology around the Himalayas is coherently reproduced by various datasets, particular attention should be paid to significant local features in the high mountain ranges of Asia. High-resolution spatial features showed that the precipitation amount around the summit of Mt. Everest is not plentiful, as reported in previous studies based on in situ observations (e.g., Dhar and Nandargi 2000; Harper and Humphrey 2003; Shea et al. 2015). The PR dataset provided quantitative information regarding the high-elevation precipitation gradients. This is potentially important to help reduce input data uncertainties for glacier nourishment (Shea et al. 2015) and to understand snow accumulation at the ridge scale (Dadic et al. 2010; Mott and Lehning 2010) by mitigating current clutter contamination and evaluating the retrieval deficiency in snowfall detection. PR has strengths when observing orographic rainfall; however, it also possesses some uncertainties, as argued in section 3b. Near-surface dense and high-sensitivity precipitation observations are required to separate out noisy climatology related to main-lobe clutter over specific areas, such as the slopes of the Himalayas.
The suppression of rainfall along the middle of the Amazon River was most distinctive in the afternoon. A nocturnal peak was detected for localized storms, as was investigated by Romatschke and Houze (2013). Sharp diurnal variations in rainfall from mesoscale storms with sizes between 10 and 100 km were identified along the course of the river. A horizontal pattern with substantial geographic constraints underlines the importance of finescale information in understanding the local climate and in grasping the spatial representativeness and uniformity of the region. Over the Brazilian Amazon, migrating widespread systems (>100 km) have a large impact on regional diurnal characteristics around the Amazon basin, except for areas along river courses. Further separations of the precipitation regimes are expected to improve our understanding of various aspects of local climatology. Continuing efforts to quantify uncertainties in finescale PR climatology are also necessary; nevertheless, results of the detected spatial gradients correlated to surface types near water bodies place greater emphasis on efforts to mitigate the deficiency of the microwave radiometers when detecting finescale features over the areas containing water bodies.
The statistical rainfall enhancement (RE) over small islands was explained in terms of island properties (area and maximum elevation) and scale-based precipitation systems. Approximately 5% of the rainfall was enhanced over very small islands (0.1°). This was particularly conspicuous for islands with high mountains and for cases with contributions from small precipitation systems. For the very small flat islands thought to have insignificant effects (e.g., Sobel et al. 2011, 2013), the RE was 1% for all PR-PSs and 26% for small PR-PSs. This implies that even for minor islands there are nonnegligible island effects of atmospheric moisture convergence. Larger islands had a positive RE because of medium-scale systems. By contrast, a negative anomaly prominently appeared for precipitation systems larger than islands. Large PR-PSs were generally enhanced in the morning over the ocean, especially for areas several tens of kilometers off the coast, resulting in a small impact on the average rainfall over tropical islands. Nevertheless, rainfall over small islands increased because of the absence of significant rainfall concentrations from large-scale systems over the coastal ocean and a marked peak from medium-scale storms over islands, indicating a possible connection with the local land–sea breeze around small islands. Interestingly, the development of widespread systems is regionally influenced by the presence of small islands. When it comes to large islands and continents in the tropics, the diurnal amplitudes of rainfall increased in extensively coastal areas. The rainfall concentration appeared over coastal oceans near continents and the inland regions of large islands. The topography and coastline curvature modulates the timing and location of precipitation (Baker et al. 2001; Qian et al. 2012). In the midlatitudes, there is a large discontinuity in the rainfall between continents and adjacent oceans. A transition and slight increase in the rainfall are seen within a few tens of kilometers of the coastline. The observational result that the cross-shore diurnal variation at latitudes greater than 30° has a markedly different pattern from that in the lower latitudes is in agreement with the classical theory of Rotunno (1983). Accumulating more data will allow our understanding to be more comprehensive. For example, the actual topographic effects on coastal rainfall and the latitudinal shifts in the diurnal timing of sea-breeze circulations at different latitudes need to be investigated further. The diurnally varying rainfall along the coastline at higher latitudes is expected to be resolved by finescale GPM DPR precipitation climatology in the future.
These results on island RE and local concentrations around the coast confirmed that some localized data do not represent the areal rainfall. Dividing total rainfall into components from scale-based storms at fine spatial and temporal scales identified significant island influences. A finescale dataset will lead to an understanding of the microprecipitation climatology over land and allow us to pursue the accurate validations or area representations required for local hydrological use (e.g., Krajewski and Ciach 2003; Villarini et al. 2008; Ward et al. 2011). Further work is needed to develop strategies for ground observations, particularly over ungauged regions and areas with significant variations in time and space, on the basis of the finescale nature of precipitation features.
This study reaffirmed long-term statistical values for assessing various global rainfall datasets, estimating nonuniform rainfall over watersheds, and describing the strength and uncertainty of spaceborne radar data. Because of local retrieval uncertainties, several difficulties remain in interpreting the pattern diversity of rainfall data. Regional rainfall estimates using microwave radiometers that require a priori information concerning the profiles, such as those of orographic rainfall, are also highly valuable. As noted, long-term TRMM PR and consecutive GPM DPR data are expected to reduce retrieval errors and allow an assessment of the credibility of climatic issues. The resolution enhancement will be of significant value from the perspective of the known retrieval error reduction of clutter contamination, incidence angle dependency, and nonuniform beam-filling effects, which will further improve our understanding of microscale rainfall features linked with ground observation networks and social needs.
Acknowledgments
The authors would like to express their gratitude to the members of the TRMM and GPM projects. Thanks are given to Mr. Tsukamoto and Mr. Okada for fruitful discussions on local climatology. The authors are grateful for the constructive comments of the anonymous reviewers. This work was primarily supported by the seventh and eighth GPM/TRMM RA of JAXA.
REFERENCES
Ageta, Y., 1976: Characteristics of precipitation during monsoon season in Khumbu Himal. Seppyo, 38, 84–88.
Andermann, C., S. Bonnet, R. Gloaguen, 2011: Evaluation of precipitation data sets along the Himalayan front. Geochem. Geophys. Geosyst., 12, Q07023, doi:10.1029/2011GC003513.
Anders, A. M., and S. W. Nesbitt, 2015: Altitudinal precipitation gradients in the tropics from Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar. J. Hydrometeor., 16, 441–448, doi:10.1175/JHM-D-14-0178.1.
Angelis, C. F., G. R. McGregor, and C. Kidd, 2004: Diurnal cycle of rainfall over the Brazilian Amazon. Climate Res., 26, 139–149, doi:10.3354/cr026139.
Baker, R. D., B. H. Lynn, A. Boone, W.-K. Tao, and J. Simpson, 2001: The influence of soil moisture, coastline curvature, and land-breeze circulations on sea-breeze-initiated precipitation. J. Hydrometeor., 2, 193–211, doi:10.1175/1525-7541(2001)002<0193:TIOSMC>2.0.CO;2.
Bergemann, M., C. Jakob, and T. P. Lane, 2015: Global detection and analysis of coastline-associated rainfall using an objective pattern recognition technique. J. Climate, 28, 7225–7236, doi:10.1175/JCLI-D-15-0098.1.
Bhatt, B. C., and K. Nakamura, 2006: A climatological–dynamical analysis associated with precipitation around the southern part of the Himalayas. J. Geophys. Res., 111, D02115, doi:10.1029/2005JD006197.
Biasutti, M., S. E. Yuter, C. D. Burleyson, and A. H. Sobel, 2012: Very high resolution rainfall patterns measured by TRMM Precipitation Radar: Seasonal and diurnal cycles. Climate Dyn., 39, 239–258, doi:10.1007/s00382-011-1146-6.
Cronin, T. W., K. A. Emanuel, and P. Molnar, 2015: Island precipitation enhancement and the diurnal cycle in radiative–convective equilibrium. Quart. J. Roy. Meteor. Soc., 141, 1017–1034, doi:10.1002/qj.2443.
Dadic, R., R. Mott, M. Lehning, and P. Burlando, 2010: Wind influence on snow depth distribution and accumulation over glaciers. J. Geophys. Res., 115, F01012, doi:10.1029/2009JF001261.
de Oliveira, A. P., and D. R. Fitzjarrald, 1993: The Amazon River breeze and the local boundary layer: I. Observations. Bound.-Layer Meteor., 63, 141–162, doi:10.1007/BF00705380.
Dhar, O. N., and S. Nandargi, 2000: An appraisal of precipitation distribution around the Everest and Kanchenjunga peaks in the Himalayas. Weather, 55, 223–234, doi:10.1002/j.1477-8696.2000.tb04065.x.
Duan, Y., A. M. Wilson, and A. P. Barros, 2015: Scoping a field experiment: Error diagnostics of TRMM Precipitation Radar estimates in complex terrain as a basis for IPHEx2014. Hydrol. Earth Syst. Sci., 19, 1501–1520, doi:10.5194/hess-19-1501-2015.
Fang, J., J. Du, W. Xu, P. Shi, M. Li, and X. Ming, 2013: Spatial downscaling of TRMM precipitation data based on the orographical effect and meteorological conditions in a mountainous area. Adv. Water Resour., 61, 42–50, doi:10.1016/j.advwatres.2013.08.011.
Farr, T. G., and Coauthors, 2007: The Shuttle Radar Topography Mission. Rev. Geophys., 45, RG2004, doi:10.1029/2005RG000183.
Fitzjarrald, D. R., R. K. Sakai, O. L. L. Moraes, R. Cosme de Oliveira, O. C. Acevedo, M. J. Czikowsky, and T. Beldini, 2008: Spatial and temporal rainfall variability near the Amazon–Tapajós confluence. J. Geophys. Res., 113, G00B11, doi:10.1029/2007JG000596.
Gille, S. T., S. G. L. Smith, and N. M. Statom, 2005: Global observations of the land breeze. Geophys. Res. Lett., 32, L05605, doi:10.1029/2004GL022139.
Grossman, R. L., and O. Garcia, 1990: The distribution of deep convection over ocean and land during the Asian summer monsoon. J. Climate, 3, 1032–1044, doi:10.1175/1520-0442(1990)003<1032:TDODCO>2.0.CO;2.
Hamada, A., and Y. N. Takayabu, 2014: A removal filter for suspicious extreme rainfall profiles in TRMM PR 2A25 version 7 data. J. Appl. Meteor. Climatol., 53, 1252–1271, doi:10.1175/JAMC-D-13-099.1.
Harper, J. T., and N. F. Humphrey, 2003: High altitude Himalayan climate inferred from glacial ice flux. Geophys. Res. Lett., 30, 1764, doi:10.1029/2003GL017329.
Heiblum, R. H., I. Koren, and O. Altaratz, 2011: Analyzing coastal precipitation using TRMM observations. Atmos. Chem. Phys., 11, 13 201–13 217, doi:10.5194/acp-11-13201-2011.
Hijmans, R. J., S. E. Cameron, J. L. Parra, P. G. Jones, and A. Jarvis, 2005: Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol., 25, 1965–1978, doi:10.1002/joc.1276.
Hirose, M., 2015: Finescale climatology of widespread precipitation systems observed by TRMM PR. Proc. 2015 IEEE Int. Geoscience and Remote Sensing Symp., Milan, Italy, IEEE, 15615146, doi:10.1109/IGARSS.2015.7326990.
Hirose, M., and K. Nakamura, 2005: Spatial and diurnal variation of precipitation systems over Asia observed by the TRMM Precipitation Radar. J. Geophys. Res., 110, D05106, doi:10.1029/2004JD004815.
Hirose, M., R. Oki, D. A. Short, and K. Nakamura, 2009: Regional characteristics of scale-based precipitation systems from ten years of TRMM PR data. J. Meteor. Soc. Japan, 87A, 353–368, doi:10.2151/jmsj.87A.353.
Hirose, M., S. Shimizu, R. Oki, T. Iguchi, D. A. Short, and K. Nakamura, 2012: Incidence-angle dependency of TRMM PR rain estimates. J. Atmos. Oceanic Technol., 29, 192–206, doi:10.1175/JTECH-D-11-00067.1.
Hirose, M., Y. N. Takayabu, A. Hamada, S. Shige, and M. K. Yamamoto, 2017: Impact of long-term observation on the sampling characteristics of TRMM PR precipitation. J. Appl. Meteor. Climatol., 56, 713–723, doi:10.1175/JAMC-D-16-0115.1.
Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement Mission. Bull. Amer. Meteor. Soc., 95, 701–722, doi:10.1175/BAMS-D-13-00164.1.
Houze, R. A., Jr., 2012: Orographic effects on precipitating clouds. Rev. Geophys., 50, RG1001, doi:10.1029/2011RG000365.
Houze, R. A., Jr., K. L. Rasmussen, M. D. Zuluaga, and S. R. Brodzik, 2015: The variable nature of convection in the tropics and subtropics: A legacy of 16 years of the Tropical Rainfall Measuring Mission satellite. Rev. Geophys., 53, 994–1021, doi:10.1002/2015RG000488.
Iguchi, T., T. Kozu, J. Kwiatkowski, R. Meneghini, J. Awaka, and K. Okamoto, 2009: Uncertainties in the rain profiling algorithm for the TRMM Precipitation Radar. J. Meteor. Soc. Japan, 87A, 1–30, doi:10.2151/jmsj.87A.1.
Immerzeel, W. W., M. F. P. Bierkens, J. Shea, A. B. Shrestha, F. Pellicciotti, and G. Rasul, 2015: Calibrating above and below snow line precipitation as inputs to mountain hydrology models—Final report. Utrecht University Rep., 205 pp.
Isotta, F. A., and Coauthors, 2014: The climate of daily precipitation in the Alps: Development and analysis of a high-resolution grid dataset from pan-Alpine rain-gauge data. Int. J. Climatol., 34, 1657–1675, doi:10.1002/joc.3794.
Janowiak, J. E., V. E. Kousky, and R. J. Joyce, 2005: Diurnal cycle of precipitation determined from the CMORPH high spatial and temporal resolution global precipitation analyses. J. Geophys. Res., 110, D23105, doi:10.1029/2005JD006156.
Kousky, V., 1980: Diurnal rainfall variation in Northeast Brazil. Mon. Wea. Rev., 108, 488–498, doi:10.1175/1520-0493(1980)108<0488:DRVINB>2.0.CO;2.
Krajewski, W. F., and G. J. Ciach, 2003: An analysis of small-scale rainfall variability in different climatic regimes. Hydrol. Sci. J., 48, 151–162, doi:10.1623/hysj.48.2.151.44694.
Liu, C., and E. J. Zipser, 2014: Differences between the surface precipitation estimates from the TRMM Precipitation Radar and Passive Microwave Radiometer version 7 products. J. Hydrometeor., 15, 2157–2175, doi:10.1175/JHM-D-14-0051.1.
Liu, C., E. J. Zipser, D. J. Cecil, S. W. Nesbitt, and S. Sherwood, 2008: A cloud and precipitation feature database from nine years of TRMM observations. J. Appl. Meteor. Climatol., 47, 2712–2728, doi:10.1175/2008JAMC1890.1.
Mair, A., and A. Fares, 2011: Comparison of rainfall interpolation methods in a mountainous region of a tropical island. J. Hydrol. Eng., 16, 371–383, doi:10.1061/(ASCE)HE.1943-5584.0000330.
Mapes, B. E., T. T. Warner, and M. Xu, 2003: Diurnal patterns of rainfall in northwestern South America. Part III: Diurnal gravity waves and nocturnal convection offshore. Mon. Wea. Rev., 131, 830–844, doi:10.1175/1520-0493(2003)131<0830:DPORIN>2.0.CO;2.
McCollum, J. R., and R. R. Ferraro, 2005: Microwave rainfall estimation over coasts. J. Atmos. Oceanic Technol., 22, 497–512, doi:10.1175/JTECH1732.1.
Mega, T., and S. Shige, 2016: Improvements of rain/no-rain classification methods for microwave radiometer over coasts by dynamic surface-type classification. J. Atmos. Oceanic Technol., 33, 1257–1270, doi:10.1175/JTECH-D-15-0127.1.
Ménégoz, M., H. Gallée, and H. W. Jacobi, 2013: Precipitation and snow cover in the Himalayas: From reanalysis to regional climate simulations. Hydrol. Earth Syst. Sci., 17, 3921–3936, doi:10.5194/hess-17-3921-2013.
Mori, S., and Coauthors, 2011: Convective systems developed along the coastline of Sumatera Island, Indonesia, observed with an X-band Doppler radar during the HARIMAU2006 campaign. J. Meteor. Soc. Japan, 89A, 61–81, doi:10.2151/jmsj.2011-A04.
Mott, R., and M. Lehning, 2010: Meteorological modeling of very high-resolution wind fields and snow deposition for mountains. J. Hydrometeor., 11, 934–949, doi:10.1175/2010JHM1216.1.
Negri, A. J., R. F. Adler, and L. Xu, 2002: A TRMM-calibrated infrared rainfall algorithm applied over Brazil. J. Geophys. Res., 107, 8048, doi:10.1029/2000JD000265.
Nesbitt, S. W., and A. M. Anders, 2009: Very high resolution precipitation climatologies from the Tropical Rainfall Measuring Mission Precipitation Radar. Geophys. Res. Lett., 36, L15815, doi:10.1029/2009GL038026.
Nullet, D., and M. McGranaghan, 1988: Rainfall enhancement over the Hawaiian Islands. J. Climate, 1, 837–839, doi:10.1175/1520-0442(1988)001<0837:REOTHI>2.0.CO;2.
Ogino, S., M. Yamanaka, S. Mori, and J. Matsumoto, 2016: How much is the precipitation amount over the tropical coastal region? J. Climate, 29, 1231–1236, doi:10.1175/JCLI-D-15-0484.1.
Palazzi, E., J. von Hardenberg, and A. Provenzale, 2013: Precipitation in the Hindu-Kush Karakoram Himalaya: Observations and future scenarios. J. Geophys. Res. Atmos., 118, 85–100, doi:10.1029/2012JD018697.
Park, N.-W., 2013: Spatial downscaling of TRMM precipitation using geostatistics and fine scale environmental variables. Adv. Meteor., 2013, 237126, doi:10.1155/2013/237126.
Qian, J.-H., 2008: Why precipitation is mostly concentrated over islands in the Maritime Continent. J. Atmos. Sci., 65, 1428–1441, doi:10.1175/2007JAS2422.1.
Qian, T., C. C. Epifanio, and F. Zhang, 2012: Topographic effects on the tropical land and sea breeze. J. Atmos. Sci., 69, 130–149, doi:10.1175/JAS-D-11-011.1.
Rasmussen, K. L., M. M. Chaplin, M. D. Zuluaga, and R. A. Houze Jr., 2016: Contribution of extreme convective storms to rainfall in South America. J. Hydrometeor., 17, 353–367, doi:10.1175/JHM-D-15-0067.1.
Roca, R., J. Aublanc, P. Chambon, T. Fiolleau, and N. Viltard, 2014: Robust observational quantification of the contribution of mesoscale convective systems to rainfall in the tropics. J. Climate, 27, 4952–4958, doi:10.1175/JCLI-D-13-00628.1.
Roe, G. H., 2005: Orographic precipitation. Annu. Rev. Earth Planet. Sci., 33, 645–671, doi:10.1146/annurev.earth.33.092203.122541.
Romatschke, U., and R. A. Houze Jr., 2011: Characteristics of precipitation convective systems in the South Asian monsoon. J. Hydrometeor., 12, 3–26, doi:10.1175/2010JHM1289.1.
Romatschke, U., and R. A. Houze Jr., 2013: Characteristics of precipitating convective systems accounting for the summer rainfall of tropical and subtropical South America. J. Hydrometeor., 14, 25–46, doi:10.1175/JHM-D-12-060.1.
Rotunno, R., 1983: On the linear theory of the land and sea breeze. J. Atmos. Sci., 40, 1999–2009, doi:10.1175/1520-0469(1983)040<1999:OTLTOT>2.0.CO;2.
Seo, E.-K., S. Hristova-Veleva, G. Liu, M.-L. Ou, and G.-H. Ryu, 2015: Long-term comparison of collocated instantaneous rain retrievals from the TRMM Microwave Imager and Precipitation Radar over the ocean. J. Appl. Meteor. Climatol., 54, 867–879, doi:10.1175/JAMC-D-14-0235.1.
Seto, S., and T. Iguchi, 2007: Rainfall-induced changes in actual surface backscattering cross sections and effects on rain-rate estimates by spaceborne precipitation radar. J. Atmos. Oceanic Technol., 24, 1693–1709, doi:10.1175/JTECH2088.1.
Shea, J. M., W. W. Immerzeel, P. Wagnon, C. Vincent, and S. Bajracharya, 2015: Modelling glacier change in the Everest region, Nepal Himalaya. Cryosphere, 9, 1105–1128, doi:10.5194/tc-9-1105-2015.
Shi, Y., and L. Song, 2015: Spatial downscaling of monthly TRMM precipitation based on EVI and other geospatial variables over the Tibetan Plateau from 2001 to 2012. Mt. Res. Dev., 35, 180–194, doi:10.1659/MRD-JOURNAL-D-14-00119.1.
Shige, S., S. Kida, H. Ashiwake, T. Kubota, and K. Aonashi, 2013: Improvement of TMI rain retrievals in mountainous areas. J. Appl. Meteor. Climatol., 52, 242–254, doi:10.1175/JAMC-D-12-074.1.
Shige, S., M. K. Yamamoto, and A. Taniguchi, 2014: Improvement of TMI rain retrieval over the Indian Subcontinent. Remote Sensing of the Terrestrial Water Cycle, Geophys. Monogr., Vol. 206, Amer. Geophys. Union, 27–42.
Shrestha, D., P. Singh, and K. Nakamura, 2012: Spatiotemporal variation of rainfall over the central Himalayan region revealed by TRMM Precipitation Radar. J. Geophys. Res., 117, D22106, doi:10.1029/2012JD018140.
Silva Dias, M. A. F., P. L. Silva Dias, M. Longo, D. R. Fitzjarrald, and A. S. Denning, 2004: River breeze circulation in eastern Amazonia: Observations and modelling results. Theor. Appl. Climatol., 78, 111–121, doi:10.1007/s00704-004-0047-6.
Sobel, A. H., C. D. Burleyson, and S. E. Yuter, 2011: Rain on small tropical islands. J. Geophys. Res., 116, D08102, doi:10.1029/2010JD014695.
Sobel, A. H., C. D. Burleyson, S. E. Yuter, and M. Biasutti, 2013: Correction to “Rain on small tropical islands.” J. Geophys. Res. Atmos., 118, 2301–2302, doi:10.1002/jgrd.50205.
Takahashi, N., and T. Iguchi, 2007: Possible improvements in the standard algorithms for TRMM/PR. 33rd Int. Conf. on Radar Meteorology, Cairns, Australia, Amer. Meteor. Soc., P3.4. [Available online at https://ams.confex.com/ams/33Radar/techprogram/paper_123569.htm.]
Takahashi, N., and R. Oki, 2010: Development of a digital elevation map using TRMM/PR data. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci, 38, 150–153.
Tartari, G., G. Verza, and L. Bertolami, 1998: Meteorological data at the Pyramid Observatory Laboratory (Khumbu Valley, Sagarmatha National Park, Nepal). Limnology of High Altitude Lakes in the Mt Everest Region (Nepal), A. Lami and G. Giussani, Eds., Memorie dell'Istituto Italiano di Idrobiologia, Vol. 57, 23–40.
Tian, Y., and C. D. Peters-Lidard, 2007: Systematic anomalies over inland water bodies in satellite-based precipitation estimates. Geophys. Res. Lett., 34, L14403, doi:10.1029/2007GL030787.
Villarini, G., P. V. Mandapaka, W. F. Krajewski, and R. J. Moore, 2008: Rainfall and sampling uncertainties: A rain gauge perspective. J. Geophys. Res., 113, D11102, doi:10.1029/2007JD009214.
Wang, J.-J., R. F. Adler, G. J. Huffman, and D. Bolvin, 2014: An updated TRMM composite climatology of tropical rainfall and its validation. J. Climate, 27, 273–284, doi:10.1175/JCLI-D-13-00331.1.
Ward, E., W. Buytaert, L. Peaver, and H. Wheater, 2011: Evaluation of precipitation products over complex mountainous terrain: A water resources perspective. Adv. Water Resour., 34, 1222–1231, doi:10.1016/j.advwatres.2011.05.007.
Williams, E., T. Chan, and D. Boccippio, 2004: Islands as miniature continents: Another look at the land–ocean lightning contrast. J. Geophys. Res., 109, D16206, doi:10.1029/2003JD003833.
Xie, S.-P., H. Xu, N. H. Saji, Y. Wang, and W. T. Liu, 2006: Role of narrow mountains in large-scale organization of Asian monsoon convection. J. Climate, 19, 3420–3429, doi:10.1175/JCLI3777.1.
Yamamoto, M. K., K. Ueno, and K. Nakamura, 2011: Comparison of satellite precipitation products with rain gauge data for the Khumb region, Nepal Himalayas. J. Meteor. Soc. Japan, 89, 597–610, doi:10.2151/jmsj.2011-601.
Yamamoto, M. K., I. Tanaka, and S. Shige, 2017: Improvement of the rain/no-rain classification method for microwave radiometers over the Tibetan Plateau. IEEE Geosci. Remote Sens. Lett., doi.org/10.1109/LGRS.2017.2666814, in press.
Yan, H., and R. A. Anthes, 1987: The effects of latitude on the sea breeze. Mon. Wea. Rev., 115, 936–956, doi:10.1175/1520-0493(1987)115<0936:TEOLOT>2.0.CO;2.
Yang, Y., and Y.-L. Chen, 2008: Effects of terrain heights and sizes on island-scale circulations and rainfall for the island of Hawaii during HaRP. Mon. Wea. Rev., 136, 120–146, doi:10.1175/2007MWR1984.1.
Yasunari, T., and J. Inoue, 1978: Characteristics of monsoonal precipitation around peaks and ridges in Shorong and Khumbu Himal. Seppyo, 40, 26–32.
Zagrodnik, J. P., and H. Jiang, 2013: Investigation of PR and TMI version 6 and version 7 rainfall algorithms in landfalling tropical cyclones relative to the NEXRAD stage-IV Multisensory Precipitation Estimate dataset. J. Appl. Meteor. Climatol., 52, 2809–2827, doi:10.1175/JAMC-D-12-0274.1.