We aim to georeference (i.e., acquire the latitude and longitude of) each dam listed in the ICOLD WRD by using the nominal location (e.g., a descriptive address for a dam or reservoir) available in the WRD attributes. Examples of the attributes that are important for georeferencing include the names of the dam and reservoir, the administrative divisions the dam is affiliated with, and the name of the impounded river. Using such attribute information, spatial coordinates of a dam may be either (a) queried from an existing register or inventory where dam records were already georeferenced and verified or (b) estimated through a geocoding service that can convert nominal locations to numeric spatial coordinates. Our preference was the former when possible to optimize the georeferencing accuracy.

The schematic procedure of GeoDAR production is illustrated in Fig. 1. We started by removing duplicate records from the ∼ 59 000 dams listed in the original ICOLD WRD (accessed in March 2019). Here “duplicates” are defined as the dams that are either (a) repeatedly recorded with identical (or highly similar) attribute information or (b) different dam structures but associated with the same reservoir. Examples of the second scenario include a reservoir's primary and secondary or auxiliary dams such as the Boonton Dam and the Parsippany Dike (40.884∘ N, 74.408∘ W) in New Jersey and multiple controls for one reservoir such as Veersedam and Zandkreekdam for Veerse Meer (51.549∘ N, 3.678∘ E) in the Netherlands. Although “duplicates” in this scenario refer to different dam bodies, including them could lead to double or multiple counting of the storage capacity of the same reservoir, and similar to the production of GRanD, our goal was to link one reservoir to one dam (if possible). After removing the identified duplicates, the cleaned WRD contains 56 815 unique dams or reservoirs. These dams have an accumulative storage capacity of 7328 km3 based on the original WRD attribute values (which are occasionally missing or erroneous) or 7720 km3 after replacement or correction by Wada et al. (2017) and GRanD (see Sect. 2.4). Unless otherwise described, the ICOLD WRD mentioned in the following text refers to the version after duplicate removal. We acknowledge that owing to the challenges of lacking explicit spatial information and occasional attribute errors in the WRD, our duplicate removal is not perfect and may have misidentified or missed some duplicate dams.

We then compared the unique ICOLD WRD records against a collection of georeferenced dam registers we acquired from regional water authorities and agencies. When the attribute information of a WRD dam matched that in a regional register, the spatial coordinates from the latter were “borrowed” to the WRD record. We term this process “geo-matching”, which resulted in the georeferencing of 13 190 WRD dams. For the remaining dams in the WRD, we applied the alternative approach, “geocoding”, which transforms a nominal location (such as the dam or reservoir address formulated by ICOLD attribute information) to a pair of spatial coordinates. The tool we used to implement geocoding was the Google Maps geocoding API (http://developers.google.com/maps, last access: 14 February 2022). The geocoding process successfully retrieved the spatial coordinates of another 9338 WRD dams. The combined output from both geo-matching and geocoding were next collated with the spatial coordinates and reservoir storage capacities of 133 WRD dams larger than 10 km3 as documented in Wada et al. (2017). These processes resulted in GeoDAR v1.0, a total of 22 560 georeferenced WRD dam points with an accumulative storage capacity of 6441 km3 (accounting for more than 80 % of that in the ICOLD WRD). The Venn diagram in Fig. 2a provides an overview of the logical relations among the georeferencing sources and methods for GeoDAR v1.0.

To further improve our spatial inventory of the world's largest dams, we performed a harmonization between the dam points in GeoDAR v1.0 and GRanD v1.3. The harmonization aimed at merging both datasets, removing duplicates in the overlapped portion between them, and when possible associating new dams supplemented by GRanD with the corresponding WRD records. This process identified another 2223 dam points, including 1414 associated with the WRD but not georeferenced in GeoDAR v1.0. With removal of duplicates, this harmonization led to a total number of 24 783 georeferenced dam points, with an accumulative storage capacity of 7384 km3. An overview of this harmonization process is illustrated by the Venn diagram in Fig. 2b. Finally, the reservoir polygons for each of the georeferenced dams were retrieved as thoroughly as possible from three global water body datasets: GRanD v1.3 reservoirs (Lehner et al., 2011), HydroLAKES v1.0 (Messager et al., 2016), and the Landsat-based UCLA Circa 2015 Lake Inventory (Sheng et al., 2016). These nearly 25 000 dam points and their associated reservoir polygons constitute GeoDAR v1.1. Details of production processes, including quality assurance and quality control (QA/QC), are included in the following method sections.

The ICOLD WRD was a collective contribution from more than 100 member nations, some of which also release detailed and publicly accessible dam registers that have been georeferenced. These regional and local registers, with reliable spatial coordinates already provided for each dam, were our preferred sources for georeferencing the WRD. Since this type of register is not available for most countries, we searched several water authority and project websites and collected seven georeferenced regional registers or inventories that are open-access. Their names, sources, and numbers of documented dams are summarized in Table 2.

These seven registers and inventories cover Brazil, Canada, the United States, 31 European countries (including part of Russia), South Africa, and part of Southeast Asia (Cambodia and Myanmar), with a total dam count of more than 126 000. Besides spatial coordinates, each of these registers also provides attributes for their documented dams, which were required by the geo-matching process. While other dam inventories could be available, our geo-matching effort for GeoDAR v1.0 was focused on these collected ones. However, we referred to additional registers or inventories from China, India, and Japan (Table 2) for the validation of our WRD geocoding (see “Validation”). For these additional regional registers, it was either inconvenient to bulk-download the dam records, or we were legally restricted from releasing their dam coordinates. Therefore, we only used these registers for the purpose of validation.

The procedure of geo-matching is illustrated in Fig. 3. Given each regional register, our goal was to find its matching records from the subset of the ICOLD WRD for the same region, by cross-checking value similarities for several key attributes between the two datasets. On one hand, the compared attributes must be mutually available in both datasets. On the other hand, the attributes should cover various themes so that in combination, they are able to disambiguate records that represent different dams but may coincide in certain attributes. Taking both requirements into account, the key attributes used include the dam and reservoir names, multiple levels of administrative or political divisions for the dam, and the dam's completion year. The river on which the dam was constructed was also considered for all regions except Cambodia as the register does not contain such an attribute. For each of the key attributes, we considered values in the WRD and the regional register agreeing with each other if the similarity score between the value sequences exceeded ∼ 85 % (meaning that there are more than eight pairs of identical elements, with consideration of their orders, between two 10-character sequences). This similarity threshold tolerated minor variations in spelling that may occur among different data sources. If an agreement was not reached between the two full sequences (e.g., “Maharashtra Pradesh” and “Maharashtra”), the similarity was then tested at the level of the main subsets of the sequences in order to increase the matching success.

One of the geo-matching challenges was that the levels of political or administrative divisions are not always comparable or consistent between the WRD and the regional registers. In the WRD, the divisions were provided at the levels of country, state/province, and the nearest town/city, which are inconsistent with some of the registers. For example, the register for Brazil (Dams Safety Report in 2017) provides the finest division at the county level, whereas the European inventory (from the MARS (Managing Aquatic ecosystems and water Resources under multiple Stress) project) documents no divisions below the national level. To improve the feasibility in division comparison, we performed a “reverse geocoding” for each georeferenced regional register using the Google Maps geocoding API. Opposite to regular (or “forward”) geocoding, which converts a nominal location to numeric spatial coordinates, this reverse geocoding converted the spatial coordinates of each dam documented in the register to a parsed address that contains administrative divisions at consecutive levels. These multi-level divisions and subdivisions were appended to the original regional registers (Fig. 3), thus enabling a more flexible and complete comparison with the WRD attributes and thus an increased success rate of geo-matching.

We considered a WRD record to match a regional record if their agreements on the key attributes warranted reasonable confidence that the two records are the same dam. In principle, high confidence would require a unanimous agreement on all key attributes. However, this ideal scenario was often unnecessary and sometimes impossible. One of the reasons is that the key attributes do not always have valid values. In the WRD, for instance, the values of “nearest town” for nearly all (> 99 %) US dams are missing. While this attribute is available for many other dams, the nearest town/city is not necessarily the division administrating or containing the dam. Another reason is that our collected multi-source datasets were not collated by a universal standard. As a result, inherent discrepancies of the attribute definitions and/or values may exist among the datasets. One example is the dam's “completion year”, which could be ambiguous between the year when the dam construction was concluded and the year when the dam operation was initiated or commissioned. These two definitions do not necessarily lead to the same year. To address such inconsistencies, we defined a baseline scenario that required any pair of matched WRD and regional records to agree on the following:

dam or reservoir name;

Following the automated QA, we performed a manual QC to reassure the accuracy of the geo-matching results. We went through each geo-matched WRD record to examine whether its attributes (e.g., dam or reservoir name, administrative locations, river name, construction year, and storage capacity) indeed agreed with those of the regional source. If an evident discrepancy was identified, the “match” was removed or corrected in the final product. Although we made every endeavor to be as rigorous as possible, remnant matching errors may still exist due to the challenges of incompleteness and possible errors in the attribute information (refer to Sect. 4 for accuracies). For occasional cases that a dam was matched correctly to the register attributes but misplaced due to poorer quality of the spatial coordinates in the register, we tried to adjust or, if possible, correct the register's spatial coordinates using the best possible resources (such as Google Maps and other open-source documents). If we were unable to observe any water infrastructure at the location of a correct match, we took a conservative action and removed the match. We admit that this might mistakenly delete some of the structures (e.g., small run-of-the-river hydropower stations, weirs, and diversions) that are too small to be visible from Google Map imagery. Our manual QC identified ∼ 4 % error in the geo-matched WRD records, most of which came from QA rank M3. After removing these errors, the geo-matching process concluded with a total of 13 190 WRD records georeferenced (Fig. 3), including 3238, 6987, and 2965 for QA ranks M1, M2, and M3, respectively (Table S1 in the Supplement). The success rate, i.e., the number of geo-matched dams as a percentage of the number of WRD records, varies from about 40 % in Southeast Asia to about 80 % in South Africa and the US (Table 2), with an overall success of 71 % in all geo-matched regions (Fig. 3).

The subset of the ICOLD WRD that was not geo-matched includes the remaining 5481 (29 %) dams in the geo-matched regions and the entire 38 144 dams in the other regions of the world (Fig. 2a). For these dams, we applied the Google Maps geocoding API, a sophisticated cloud-based geocoding service, to retrieve the spatial coordinates of each dam as thoroughly and accurately as possible. To do so, we designed a recursive geocoding procedure that implemented three primary steps on each dam: forward geocoding, reverse geocoding, and QA filtering. The purpose of each of the steps and their logical relations are illustrated in Fig. 4.

The forward geocoding (see Sect. 2.1 for definition) used the text address of each dam as the input, which we formatted by concatenating the WRD attribute values, to output the latitude and longitude of the dam. The WRD attributes used for address formatting include dam name, reservoir name, state/province, and country. “Nearest town” was excluded because it is not always the township administrating the dam or reservoir. Together with the spatial coordinates, the forward geocoding also returned a Google Maps address associated with the coordinates, which was parsed to individual components including feature name, street name, and political divisions. These output address components, in return, provided valuable information for QA: if the geocoded coordinates are correct, the associated output address components should agree well with those of the WRD input. However, we noticed that address components from forwarding geocoding are often limited in terms of division levels. To complement this limitation, we also utilized reverse geocoding (see Sect. 2.2 for definition) to convert the coordinates from forward geocoding to an updated address (sometimes with more complete division levels). The address components from both forward and reverse geocoding were combined and are hereafter referred to as the “output address”.

Similar to geo-matching, we employed a QA filter to approach the optimal geocoding result. This process first arranged the attributes of each WRD record to several address formats as they could result in different geocoding outputs. The address arrangements are listed in Table S2 in the Supplement, and their preference order is rationalized in the Supplement. Each of these WRD addresses was used iteratively for both forward and reverse geocoding (as described above). Their geocoded spatial coordinates were then ranked to five discrete QA levels based on how well the input and output addresses agree with each other (C1 to C5 in Table S3 in the Supplement). The iteration could be terminated if the highest QA rank was achieved; otherwise, the coordinates that rendered the best possible QA rank were used as the geocoding result.

As explained in Table S3 in the Supplement, the compared address components include the name of the feature and its affiliated political divisions from town/city to country levels. Consistent with geo-matching, we considered a component to be agreed on if the similarity of its values from both input and output addresses exceeds ∼ 85 %. Since the nearest town in the WRD was not used for forward geocoding, we treated it as an “independent reference” for validating the township component in the output address. Although the town or city near the dam (from the WRD) does not always coincide with that administrating the dam (from the geocoding output), their occasional agreement would strengthen our confidence of the geocoded coordinates if other components were also well matched between the WRD input and the geocoding output. For this reason, we opted to include the township comparison as a supplementary criterion in the geocoding QA process. The highest QA rank (C1) corresponds to a unanimous agreement on all address components. However, the minimum rank (C5) only required the agreement on the feature name, which is a more flexible baseline in comparison with that for geo-matching. This was because some of the large reservoirs, particularly those on or near political boundaries, have shared or ambiguous divisions, and the ambiguity might be further amplified by the output coordinates, which could fall in anywhere from the dam to across the reservoir water surface. In addition, some of the outputs, regardless of agreement on the address components, are not dams or reservoirs. We therefore included another baseline filter which aimed to remove such errors by analyzing the feature type information in the geocoding output (see scripts in “Code availability”). Although the QA process was designed to be automated, we still manually enforced hundreds of the initial outputs, many of which had returned feature names in native languages, to pass the baseline filters. As a result, our QA process yielded more than 16 000 geocoded WRD records, each with the optimal spatial coordinates and the corresponding QA rank.

To complement the QA process, we then conducted a rigorous QC to correct and/or remove the remaining geocoding errors. We considered a geocoding error to be a location where (a) no dam or reservoir could be visibly verified from Google Earth or Esri images, or (b) the WRD attribute information is inconsistent with the feature or division labels on Google Maps. In such cases, we usually first attempted to re-geocode the dam manually (such as by directly using the Google Maps interface) before deleting this error. It is important to clarify that the georeferenced coordinates, although referred to as ”dam points” in our data product, do not always fall on the dam bodies. While the geo-matched coordinates from regional registers are usually on or close to the dams, the geocoded coordinates, depending on the address input (Table S2 in the Supplement) and the available Google Maps information, could be located on the associated reservoir. Note that the latter case was not considered an error, but for improved locations, we manually adjusted some of the georeferenced coordinates more towards the dams. Due to China's GPS shift problem (e.g., misalignment between the street maps and satellite imagery on Google Maps), the geocoded points across mainland China often exhibit systematic offsets of roughly 500 m or more from their actual dam or reservoir features. For such Chinese dams, we tried to reduce their geocoding offsets by manually relocating the coordinate points to their correct dams or reservoirs. Our QC process ended up removing about 42 % of the originally geocoded dams, most of which stemmed from relatively low QA ranks (see statistics in Table S3 in the Supplement). The complete geocoding procedure resulted in 9338 georeferenced and quality-controlled WRD records, with an overall success rate of 21 %.

The outputs from both geo-matching and geocoding, a total of 22 528 georeferenced ICOLD WRD records (Fig. 2a), were further supplemented or harmonized by two global dam or reservoir inventories to improve our inclusion of the world's largest dams. We considered this process necessary for two reasons. First, our georeferencing process, particularly geocoding via Google Maps API, did not warrant an exhaustive inclusion of the largest dams. This is particularly evident for regions where the address and label information in Google Maps is either lacking or difficult to pass the automated QA due to language ambiguity or naming discrepancies. Second, through cross-referencing we noted that the attribute values of reservoir storage capacity (as well as reservoir surface area) provided in the ICOLD WRD are occasionally erroneous (also noted by Mulligan et al., 2020), e.g., by a factor of 1000, probably caused by unit confusion in WRD compilation. As part of the supplementation and harmonization process, we also reduced the errors in WRD by verifying storage capacities of some of the largest reservoirs and replacing the WRD capacity values by those of the two global inventories.

Reservoir polygons of the georeferenced dam points were retrieved as thoroughly as possible from three global water body datasets: GRanD reservoirs (Lehner et al., 2011), HydroLAKES v1.0 (Messager et al., 2016), and UCLA Circa 2015 Lake Inventory (Sheng et al., 2016). These three water body datasets exhibit an increasing spatial resolution: from 7000+ polygons in GRanD reservoirs provided exclusively for GRanD's dam points to millions of water body polygons, including both natural lakes and reservoirs, in the other two datasets. While HydroLAKES documents 1.4 million water bodies larger than 0.1 km2 (10 ha), the Landsat-based UCLA Circa 2015 Lake Inventory further reduced the minimum size to only 0.004 km2 (0.4 ha), resulting in another 7.7 million water bodies on the global continental surface. Accordingly, we implemented a hierarchical procedure, where the three water body datasets were applied in ascending order of spatial resolution to retrieve the reservoir boundaries with an overall decreasing size.

Specifically, GRanD v1.3 provides 7162 valid reservoir polygons for the 7303 dam points (after coordinate corrections) used for harmonization. These GRanD polygons were first assigned to their associated dam points in GeoDAR v1.1 through GRanD IDs. Reservoirs of the remaining 17 556 dam points in GeoDAR v1.1, including the 76 GRanD dams with corrected locations (Table S5 in the Supplement), were next retrieved from HydroLAKES when possible. To avoid duplicates in the reservoirs retrieved from different data sources, we only used the subset of HydroLAKES that is spatially independent from (i.e., not intersecting with) GRanD reservoirs. Different from reservoir assignment using GRanD, there was no common attribute ID to pair HydroLAKES polygons with the remaining dam points, so their reservoir retrieval relied completely on spatial association. One major challenge in dam–reservoir spatial association was the ambiguity caused by the offsets between our georeferenced dam points and their actual reservoir polygons (see Sect. 2.3).

To tackle this challenge, we designed a procedure containing three rounds of iteration to progressively optimize reservoir-dam association. This procedure was based on two assumptions, both conditional on a reasonable spatial tolerance. We started with 500 m to be roughly consistent with the Google street map offsets for China. The first assumption was that larger reservoirs are more likely to be documented than smaller ones, in both the ICOLD WRD and Google Maps. Therefore, the first round of iteration assigned each of the dams to the largest water body within the tolerance. This assignment might, however, lead to a situation where multiple dams were assigned to the same reservoir. To untangle this situation, the remaining iterations assumed Tobler's first law of geography (Tobler, 1970): “everything is related to everything else, but near things are more related than distant things” (p. 236). Accordingly, for any water body mistakenly associated with multiple dams, the second round of iteration reassigned the water body to its closest dam, and the other dam(s) within the tolerance, as a result, was/were left unpaired. To reduce the number of such “orphan” dams, a final, third round of iteration assigned the remaining unpaired dams to the next closest water body that was within the spatial tolerance and had not been previously associated with any dams. If this led to multiple dams associated with one reservoir again, only the dam with the closest proximity to the reservoir was kept. Through experimentation, we opted to implement this three-iteration procedure twice, first using a conservative 500 m tolerance to maximize the accuracy for most associations and then a 1 km tolerance to further minimize the number of orphan dams.

This multi-iteration procedure retrieved roughly 7600 reservoir polygons from HydroLAKES. For the remaining dam points left unpaired, we applied the same association procedure to continue retrieving their reservoirs from the high-resolution UCLA Circa 2015 Lake Inventory. Similarly, only the subset that does not intersect with the retrieved HydroLAKES polygons was considered in order to avoid duplicates in the retrieved reservoirs from different datasets. The use of the UCLA Circa 2015 Lake Inventory retrieved another 6700 or so reservoirs.

We followed the automated reservoir retrieval by a manual QC to visually confirm that each retrieved reservoir polygon was matched to the correct dam point, and if not, we corrected the association as thoroughly as possible. This visual QC was particularly necessary for lake-dense regions, including the case of cascade reservoirs immediately downstream or upstream to each other. While some of the dams, such as barrages, diversion infrastructure, and dams under construction, do not have visible impoundments (Lehner et al., 2011), we tried to be as meticulous as possible to verify the orphan dams and recover any missing reservoirs. For instance, we were able to manually retrieve 10 reservoirs (including 4 completed after 2000) from the UCLA Circa 2015 Lake Inventory for the ∼ 70 dams in GRanD v1.3 without reservoir polygons. We also assigned reservoirs to 68 of the 76 GRanD dams with our corrected spatial coordinates. Although no new reservoirs were digitized (all original polygons retrieved from the three water masks), we modified the geometries of some of the reservoirs when necessary. For example, we truncated or split a polygon if we saw its original extent intruding to another reservoir and dissolved several polygons into a multipart feature if they cover the same reservoir surface. We also replaced hundreds of reservoirs initially retrieved from GRanD and HydroLAKES by the polygons in the UCLA inventory to improve the boundary accuracy and completeness.

GeoDAR v1.0 is a collection of 22 560 dam points georeferenced exclusively for the ICOLD WRD (Fig. 6a). Among them, 13 149 or 58 % were retrieved from geo-matching regional dam registers, 9278 or 41 % from Google Maps geocoding API, and the remaining 133 largest dams from the spatial inventory in Wada et al. (2017) (Fig. 6b). WRD storage capacities of most of these 133 large reservoirs were replaced by the values in Wada et al. (2017) (see Sect. 2.4.1), and unless stated otherwise, our following statistics on storage capacities were calculated after this replacement.

The total reservoir storage capacity of these dams is 6441 km3, meaning that GeoDAR v1.0 georeferenced 40 % of the 56 815 WRD records but included more than 80 % of their cumulative reservoir storage capacity. The total storage capacity of the 133 largest dams from Wada et al. (2017), despite being limited in number, reaches 3900 km3 or 61 % of the cumulative storage capacity in GeoDAR v1.0, and the other ∼ 40 % capacity was split almost equally between the remaining 22 000+ geo-matched and geocoded dams. Although the registers used for geo-matching are regional, the dams in GeoDAR v1.0, as shown in Fig. 6b, are distributed in 151 out of the 165 countries or territories in the WRD, largely owing to our geocoding efforts through Google Maps API. Since the production of v1.0 was largely independent of other global dam datasets such as GRanD, it can also be used to cross-compare, supplement, and potentially improve other dam datasets. Validation of our georeferencing accuracy for v1.0 is provided in Sect. 4.

GeoDAR v1.1 consists of (a) 24 783 dam points (Fig. 6a) representing a full harmonization between GeoDAR v1.0 and GRanD v1.3 and (b) 21 515 reservoir polygons (Fig. 7) based on a one-dam-to-one-reservoir relationship. In these nearly 25 000 dam points, 17 480 or 71 % come from GeoDAR v1.0 alone, 6494 or 26 % are shared by the ICOLD WRD and GRanD, and the other 809 or 3 % are from GRanD alone (Table 1, Fig. 6c). Among the 6494 shared dams, 5080 were georeferenced in both GeoDAR v1.0 and GRanD, and the remaining 1414 were introduced through the harmonization with GRanD. This resulted in a total of 23 974 georeferenced WRD records (42 % of all WRD records) in GeoDAR v1.1. In addition to the expanded number of georeferenced WRD dams, GRanD supplemented another 809 dams which are exclusive of the WRD. The total 2223 dams added by GRanD, notated as “GRanD v1.3 & other ICOLD” and “GRanD v1.3 only” in Fig. 6c, are distributed worldwide and complement v1.0, particularly in regions such as Africa and central Asia, where geocoding using Google Maps was challenging. After this ICOLD–GRanD harmonization, the spatial coverage of the dam points in GeoDAR v1.1 increased to 155 out of the 165 countries in the WRD (also see Table S6 in the Supplement).

As described in Sect. 2.4.2, we substituted the reservoir storage capacities in GRanD for the original capacity values of their overlapping WRD dams. As a result, the total reservoir storage capacity in GeoDAR v1.1 reaches 7384 km3, which compares to ∼ 95 % of the cumulative capacity in the entire ICOLD WRD (see Sect. 5.1 for more comparisons with ICOLD). As reported in Table 1, 81 % (6006 km3) of the total storage capacity in GeoDAR v1.1 is explained by the 5080 large dams georeferenced in both GeoDAR v1.0 and GRanD. The 17 480 smaller dams from GeoDAR v1.0 alone contribute only 7 % (507 km3) of the total storage capacity, which is roughly comparable to the subset from GRanD alone (268 km3) or the subset from GRanD and other ICOLD records (603 km3). These capacity contributions suggest that compared to GRanD, the major improvement of GeoDAR lies in the increased number of relatively small dams rather than the increase in total storage capacity of the dams (see Sect. 5.2 for more comparisons with GRanD).

Different from GeoDAR v1.0, version 1.1 also includes reservoir polygons for 21 515 or 87 % of the georeferenced dam points (Fig. 7). Reservoir polygons for the remaining 13 % of the dam points were retrieved unsuccessfully due to a combination of factors, including limited spatial resolutions of the applied water masks, missing water occurrence in the masks (when the reservoir water levels are too low), and the fact that some of the dams have no evident water impoundments at all. Nevertheless, the retrieved reservoir polygons have a cumulative area of 496 314 km2, accounting for 98 % of the total reservoir area of all georeferenced dams in GeoDAR v1.1 (reservoir areas without polygons are based on documented attributes). These retrieved reservoirs correspond to a cumulative storage capacity of 7216 km3, also accounting for nearly 98 % of the total storage capacity in v1.1. These statistics indicate that the reservoirs whose boundaries were retrieved unsuccessfully were mostly small in area and storage.

The numbers of reservoir polygons retrieved from each of the three water body datasets are comparable (about 7100–7200 each), but the total reservoir storage capacity and area generally decrease with the increasing spatial resolution of the water body datasets (Table 1). As a result, the mean reservoir polygon size decreased from 63 km2 for those retrieved from GRanD to 2 km2 from HydroLAKES and 5 km2 from the UCLA Circa 2015 Lake Inventory. This result is overall consistent with the design of our hierarchical procedure (Sect. 2.5), where smaller reservoirs were successively retrieved with the help of finer water masks. It is important to note that the retrieved polygons do not always represent the maximum water extents of the reservoirs because water boundaries in the retrieval sources were not necessarily mapped in the maximum inundation periods. For example, the UCLA Circa 2015 Lake Inventory was produced using Landsat images acquired during “lake-steady” periods (Lyons and Sheng, 2018) and thus represents the average seasonal extent of each water body (Sheng et al., 2016). Despite not always being the largest water extents, our retrieved reservoir polygons enhanced the spatial details of global reservoir locations, using which users can further expand or refine the water boundaries to their specific needs.

The GeoDAR dataset, including dam points for v1.0 and both dam points and reservoir polygons for v1.1, is provided as three separate shapefiles. For user convenience, we also duplicated the two dam point shapefiles in the comma-separated values (csv) format. The file names and attributes are explained in Table 3. Although most of our dam points were georeferenced using WRD records, our published GeoDAR complies with the legal codes of ICOLD and does not directly release any attribute from the WRD. The attributes we provide in GeoDAR, as listed in Table 3, are limited to our georeferencing methods, QA/QC, validation, and other information (such as spatial coordinates and part of the reservoir storage capacities) that is already open-source or has been permitted for use by the original producers.

Although WRD attributes are not directly available in GeoDAR, we suggest two possible ways for users to acquire at least some of the essential attributes. Upon the user's reasonable request and on a case-by-case basis, we may provide assistance in decrypting the association between GeoDAR IDs (Table 3) and ICOLD's international codes, and using the international codes, the user can link each of the dam or reservoir features in GeoDAR to the entire 40 or so proprietary attributes in the WRD. This is also based on the premise that the user needs to acquire the WRD attribute data from ICOLD themselves and that the user agrees not to release the GeoDAR–WRD association or the WRD attributes to the public. Alternatively, since we imposed no usage restrictions on our spatial features (dam points and reservoir polygons), users are free to integrate them with other datasets and tools, such as remote sensing observations and modeling, to acquire the needed attributes, particularly those not yet documented in the ICOLD WRD. Acquisition methods have been exemplified for at least the following attributes: reservoir hypsometry and bathymetry (Li et al., 2020; Yigzaw et al., 2018); surface evaporation loss (Mady et al., 2020; Zhan et al., 2019; Zhao and Gao, 2019a); operation rules (Shin et al., 2019; Yassin et al., 2019); completion years (Zhang et al., 2019); storage capacities (Liu et al., 2020); and the changes in water area (Pekel et al., 2016; Yao et al., 2019; Zhao and Gao, 2019b), level (Crétaux et al., 2011; Schwatke et al., 2015), and storage or volume (Busker et al., 2019; Crétaux et al., 2016; Gao et al., 2012; Zhang et al., 2014).

Despite our efforts to integrate multi-source registers and the Google Maps geocoding API, georeferencing the ICOLD WRD, particularly smaller dams in poorly documented regions, has proven to be challenging. This challenge was reflected by the proportion of the WRD that was spatially resolved in GeoDAR. As compared in Table 5, GeoDAR v1.0 included 40 % of the 56 815 records in the entire WRD. Although limited in number, these georeferenced dams were a compromise between geocoding quality and thoroughness (see Sect. 2.2 and 2.3) and account for ∼ 84 % of the total reservoir storage capacity in the WRD. The larger proportion in terms of storage capacity indicates that most of the sizable dams in the WRD have been spatially resolved. This message is also corroborated by Fig. 9. Nearly 70 % of the 12 412 WRD dams larger than 10 mcm, for example, have been georeferenced in GeoDAR v1.0 (Fig. 9a). While 80 % of the 21 849 WRD dams smaller than 1 mcm were not georeferenced, these smaller dams account for just 1 % of the total WRD storage capacity (Fig. 9b). After harmonization with GRanD, the proportion of the WRD georeferenced in GeoDAR v1.1 increased to 42 % by count or 92 % by storage capacity (Table 5), and these percentages represent our best result for georeferencing the WRD. By absorbing the remaining dams in GRanD as well, v1.1 has a total dam count equivalent to 44 % of the WRD and a cumulative storage capacity less than 5 % below that of the full WRD (Table 5, Fig. 9b). Compared to v1.0, the margin between the distribution curves of GeoDAR v1.1 and the WRD, particularly for relatively large dams, was further reduced (Fig. 9a). As a result, the number of dams larger than 10 mcm in GeoDAR v1.1 exceeds 80 % of that in the WRD, and the number of dams larger than 1 mcm reaches 60 % of that in the WRD.

The spatial coverage of GeoDAR, in comparison with the WRD, was summarized for each of the 165 countries with registered WRD records (Fig. 10). Our comparison focused on GeoDAR v1.1 as it represents an improved version of our spatial dam inventory. Among these 165 countries, the median of GeoDAR v1.1 coverage by dam count, i.e., the number of dams in GeoDAR v1.1 as a proportion of the number of dams in the WRD, is 62 %, with the first and third quartiles being 35 % and 89 %, respectively. As shown in Fig. 10a, better coverages tend to occur in North America, Europe, Russia, Australia, and part of South America and Africa, whereas poorer coverages are seen in East Asia, South Asia, and part of the Middle East. The coverages in China and India, for example, are only about 22 %–26 % due to a large quantity of WRD records for these two countries (23 749 in China excluding Taiwan and 5074 in India) but relatively limited information on Google Maps. Compared with dam counts, GeoDAR's coverage for reservoir storage capacity is higher overall (Fig. 10b). Among the 158 countries with documented reservoir storage capacities, the median coverage in GeoDAR reaches 98 %, with the first and third quartiles being 87 % and 100 %, respectively. If we exclude the 809 dams supplemented by GRanD alone and only consider the WRD portion of GeoDAR v1.1, the coverage becomes overall lower but by a limited extent (Fig. S1 in the Supplement). Among these countries, the median coverage of WRD dams by GeoDAR v1.1 is 59 % (with 33 % and 83 % as the first and third quartiles) in terms of dam count and 96 % (85 % and over 99 % as first and third quartiles) in terms of reservoir storage capacity, suggesting that a substantial proportion of the WRD had been georeferenced in many countries before the additional supplementation from GRanD. More detailed comparisons (among ICOLD, GranD v1.3, and GeoDAR v1.3) for each of the 165 countries are given in Table S6 in the Supplement.

Catchment areas of the reservoirs often indicate the stream order of the impounded river and thus the scales of flow and sediment alterations by the dam. Locating dams with an improved representation of catchment areas, particularly smaller ones, has been increasingly needed by hydrologic modeling and watershed management (Grill et al., 2019; Lin et al., 2019). To evaluate how GeoDAR spatially resolved the WRD in this aspect, we directly used the values of “catchment area” provided in the attributes. As many records in the WRD are missing catchment areas, we combined the available values in both the WRD and GRanD, and when a dam has catchment areas in both datasets, we preferred the value in GRanD. As reported in Table 5, the subset of the WRD georeferenced in GeoDAR v1.1 has a total catchment area of 140×106 km2, which covers 94 % of the total catchment area in the WRD. The remaining 6 % gap was largely closed by the inclusion of the remaining non-WRD dams from GRanD. It is worth mentioning that these statistics do not take into account the dams without documented catchment areas. While it is possible to retrieve catchment boundaries for GeoDAR dams (e.g., using DEM as per Mulligan et al., 2020), acquiring accurate catchment areas of the other WRD dams (which have not been georeferenced) is prohibited due to unknown locations. Therefore, our comparison was only based on the attribute values that are already available. This explains why GeoDAR georeferenced fewer than half of the WRD records by count but included more than 90 % of the total catchment area. Similar to the pattern of reservoir storage capacity, higher proportions of the WRD catchment area covered by GeoDAR are skewed towards the dams with larger catchment areas (Fig. 11a). For example, the number of dams with a catchment area larger than 10 km2 in GeoDAR equals 89 % of that in the WRD, and the coverage increases to 95 % for the dams with a catchment area larger than 100 km2.

Although GeoDAR does not include reservoir catchment boundaries, it does provide reservoir polygons for 87 % of the georeferenced dam points. As reported in Sect. 3.2, the remaining 13 % of the dam points without reservoir polygons, if inferred from their available attribute values, yield a reservoir area that is only ∼ 2 % of the total reservoir area of all GeoDAR dams. For this reason, we focus on the retrieved reservoir polygons for comparing how GeoDAR v1.1 represents the reservoir areas in the entire ICOLD WRD. Among the 21 515 polygons, 20 718 (96 %) are associated with georeferenced WRD dams. These retrieved WRD reservoirs have a total area of 476 603 km2, accounting for 90 % of the cumulative reservoir area in the WRD (Table 5). After supplementation of the remaining polygons (for dams in GRanD alone), the total reservoir area reached 496 314 km2, equivalent to 93 % of the cumulative reservoir area in the WRD. Like other attributes, the values of reservoir area are not always available in WRD records. If a WRD record is missing its area attribute value but has a reservoir polygon, we used the area of the reservoir polygon as the de facto reservoir area in calculating WRD statistics (see note for Table 5), and the other WRD records still missing reservoir areas probably contribute a minuscule fraction of the aggregated area. This way, we are essentially comparing the areas of the mapped reservoir polygons in GeoDAR v1.1 with the documented reservoir areas in the WRD. Since our retrieved reservoir polygons are not always at the maximum inundation extents, the comparison includes the uncertainties due to water mapping and errors in the WRD attributes. If we replaced the attribute reservoir areas by our polygon areas, the coverage increased from 93 % to nearly 96 % (dotted curve in Fig. 11b), indicating a global mean bias (underestimation) of about 2 %–3 % in our reservoir polygons. Keeping these limitations and uncertainties in mind, we showed in the distribution curves (Fig. 11b) that the number of GeoDAR reservoir polygons accounts for 68 % of all WRD records that have reservoir area values (either documented or de facto), and consistent with the distributions of other attributes, higher coverages for reservoir area tend to occur for larger reservoirs. For example, GeoDAR retrieved 8263 reservoirs larger than 1 km2, which account for 73 %–80 % of those in the WRD. The coverage increases to 87 %–92 % for reservoirs larger than 10 km2, although the reservoir polygon number decreases to 2570.

While GRanD emphasized dams larger than 100 mcm (or 0.1 km3), GeoDAR aimed to georeference WRD records which, by definition, have a minimum storage capacity of 3 mcm or smaller if the dam is higher than 15 m (see Sect. 1). This reduced storage threshold entailed a substantial increase in the dam quantity in GeoDAR. As compared in Table 5, GeoDAR v1.0, which was generated independently from GRanD, is already more than triple the dam quantity in GRanD (7320) and accounts for 95 % of the total reservoir storage capacity in GRanD (6881 Gt). With the harmonization with GRanD, the number of dams in GeoDAR v1.1 reaches 339 % of that in GRanD, with a total reservoir storage capacity also exceeding 7 % of that in GRanD. This comparison suggests that the improvement of GeoDAR is mainly manifested as the increased dam quantity rather than reservoir storage capacity.

The increased dam quantity in GeoDAR is manifested as a ubiquitous improvement of the spatial density of smaller dams worldwide (Fig. 12). Since GeoDAR v1.1 has absorbed GRanD v1.3, the global patterns for capacious reservoirs are overall similar between the two datasets. What is noticeably different are the proliferated density of thousands of smaller reservoirs, particularly those beyond the main focus of GRanD (such as those smaller than 100 mcm). The substantial increase in smaller dams and reservoirs is corroborated by the distribution curves in Fig. 9a, where the mode storage capacity (i.e., the capacity corresponding to the peak frequency) shifted from about 100 mcm in GRanD to about 3–5 mcm in GeoDAR (both v1.0 and v1.1). The area between the distribution curves is largely explained by the addition of ∼ 16 500 dams smaller than 100 mcm in GeoDAR v1.1 (Fig. 9a), which correspond to a total storage increase of 124 Gt or 95 % of the total storage of the dams smaller than 100 mcm in GRanD (Fig. 9b). It is important to note that the added reservoirs in GeoDAR still comply with ICOLD's definition of “large dams” (see Sect. 1). Although their aggregated storage is limited, these relatively small reservoirs are geographically widespread, meaning that they are locally significant for filling service gaps between more sporadic larger dams. Examples include hundreds of smaller dams and reservoirs that provide irrigation from southern Europe (Fig. 13b) to northwestern and central India (Fig. 13c), hydropower and water usage in central and southern China (Fig. 13a), and flood controls across the Mississippi River basin and southern Texas in the US (Fig. 13d). The sheer number of these added smaller dams and reservoirs accentuate the benefits of an improved knowledge of their spatial locations, such as what GeoDAR offers, for strategizing water and energy management and assessing fragmentation of the river ecosystems (Belletti et al., 2020; Grill et al., 2019).

To assist regional applications, we further aggregated the improvements of GeoDAR over GRanD into national scales. As shown in Fig. 14, GeoDAR's improvements in either dam count or reservoir storage capacity pervade more than 120 countries, occupying 86 % of the continental landmass (excluding Antarctica). The increase in dam count occurs in 127 out of the 155 GeoDAR countries (Fig. 14a, Table S6 in the Supplement). These countries include 18 countries without GRanD records at all (such as Haiti, United Arab Emirates, Yemen, and Bhutan), and the other 109 countries comprise 80 % of the 137 countries with GRanD records. There are slightly fewer countries with a confirmed increase in reservoir storage capacity (Fig. 14b) because some of the added WRD records are missing storage capacity values. The number of these countries is 117, including 15 without GRanD records at all.

While GeoDAR's improvements are widespread, the improvement levels are not geographically uniform (Fig. 14). Globally speaking, the spatial patterns of number and capacity increases are overall consistent, with the major hotspots concurring with large or industrialized nations (e.g., US, China, Brazil, India, and European countries) and less impressive increases in smaller, drier, and/or less developed nations (e.g., part of Africa and South America). This is reasonable as bigger and/or more developed nations usually possess a larger quantity of dam infrastructures and thus a greater potential for GeoDAR to improve. However, this pattern also reflects the disparities due to other factors, such as a possible bias in the WRD (as it is a volunteered dataset, and not all member nations contributed equally), the accessibility of regional registers for geo-matching, and geocoding challenges for different regions. The top five countries in terms of dam count increase are the US (an increase of 6039 or 314 %), China (4352 or 474 %), India (963 or 290 %), South Africa (667 or 248 %), and Spain (575 or 219 %) (Table S6 in the Supplement). These five countries cover 72 % of the global dam count increase (17 463). Similarly, the top five countries in terms of storage capacity increase are the US (123 km3 or 16 %), Canada (73 km3 or 8 %), Brazil (66 km3 or 12 %), China (44 km3 or 7 %), and India (33 km3 or 12 %), which together comprise 68 % of the global storage capacity increase (503 km3).

Certain regions with limited increases in dam count, such as the Middle East, Southeast Asia, and southern Africa, show more pronounced improvements in storage capacity. This contrast indicates that, in addition to smaller dams and reservoirs (e.g., < 100 mcm), GeoDAR also supplemented GRanD by including more capacious reservoirs. Examples are Dau Tieng Dam in Vietnam (storage capacity 1580 mcm; location 11.323∘ N, 106.341∘ E), San Roque Dam in the Philippines (990 mcm; 16.147∘ N, 120.685∘ E), Mrica Dam in Indonesia (193 mcm; 7.392∘ S, 109.605∘ E), Marib Dam in Yemen (398 mcm; 15.396∘ N, 45.244∘ E), and the recently completed Lauca Dam in Angola (5482 mcm; 9.739∘ S, 15.127∘ E). GeoDAR also inventoried some large hydroelectric projects that are under construction or consideration. Examples are Bakhtiari Dam in Iran (expected 4845 mcm; 32.958∘ N, 48.761∘ E), Bekhme Dam in Iraq (17 000 mcm; 36.701∘ N, 44.271∘ E), Diamer-Bhasha Dam in Pakistan (10 000 mcm; 35.521∘ N, 73.739∘ E), and Myitsone Dam in Myanmar (13 282 mcm; 25.691∘ N, 97.516∘ E).

By further aggregating national statistics to each continent, Fig. 14 echoes the fact that GeoDAR's major improvement lies in the quantity or spatial density of the dams rather than their total reservoir storage capacity. However, this should not overshadow the fact that improvements of both dam count and storage capacity do exist in all continents. As summarized in Fig. 14a, the continental improvement ascends from 170 more dams with a 7 km3 total capacity in Oceania to a scale of 6000–7000 more dams with a 100–200 km3 capacity in North America or Asia. Because the total storage capacity is disproportionally dominated by the largest reservoirs, and GRanD has already included most of them, the added storage capacity by GeoDAR relative to what has existed in GRanD appears limited and descends from 9 %–12 % in Asia and North America and 7 %–8 % in Oceania and South America to 1 %–3 % in Africa and Europe. By contrast, GeoDAR's dam quantity ranges from being almost double that of GRanD in Oceania and Africa to being triple to quadruple in the other continents.

A derivative benefit of the increased dam quantity is a more complete representation of the regulated watersheds, which is critical to improving discharge estimates. As revealed by the distribution curves in Fig. 11a, GeoDAR improved GRanD in the inclusion of reservoir catchment areas in two aspects. First, the exceedance of the number of reservoir catchments is almost unanimous on all area levels. This corresponds to a total increase in catchment area by 32×106 km2 or 27 % (Table 5). Second, the increase in reservoir catchments is skewed towards smaller catchments, signifying a more realistic inventory of human water regulations in the basins of lower stream orders or closer to stream headwaters. As shown in the distribution curves (Fig. 11a), the average increasing rate is augmented from about 30 % for catchments larger than 1000 km2 and about 80 % for catchments between 10 and 1000 km2 to nearly 600 % for those smaller than 10 km2. The mode of catchment areas decreases from about 200–400 km2 in GRanD to 30–100 km2 in GeoDAR, with the latter much closer to the mode of the entire WRD (15–50 km2). As a result, the number of dams with a catchment size smaller than 25 km2, for example, which is the channelization threshold for the high-resolution MERIT Basins hydrography dataset (Lin et al., 2019; Yamazaki et al., 2017), is 3570 or 27 % in GeoDAR in comparison to 695 or 10 % in GRanD. These small-catchment dams, once integrated into river networks, may substantially improve the performance of routing models. Consistent with our comparison with the WRD (Sect. 5.1), these statistics are only based on the records with valid catchment areas. Considering that missing values more likely occur for dams with smaller catchments, our reported improvement could be theoretically conservative.

The increased dam count in GeoDAR also enabled the retrieval of surface extents of another 14 000 or so smaller reservoirs (Fig. 7). The added reservoir polygons, including 10 for the dams in GRanD that originally had no reservoirs and 14 275 for GeoDAR v1.0 alone, have an average size of 1.4 km2 in comparison to 65 km3 in GRanD. They aggregate to a total area of 19 880 km2, a scale comparable to 30 Lake Meads. Although this area increase may appear substantial, it only expanded the global reservoir area in GRanD by a marginal proportion of 4 %. Similar to the pattern of storage capacities, reservoir areas follow a quasi-Pareto distribution, meaning that smaller reservoirs tend to dominate the population (or number), whereas larger reservoirs dominate the area and storage. This explains why the increase in relative area is small, but the increase in absolute quantity is double that of the entire reservoir polygons in GRanD. For example, 95 % of the total reservoir area in GeoDAR comes from only 12 % of the reservoir polygons larger than 10 km2, and about 90 % of these large reservoirs are already included by GRanD (Fig. 11b). This pattern again suggests that the core value of GeoDAR is not to augment the global scale of reservoir area or storage but to amplify the local details of smaller dams and reservoirs. Owing to the added details, the mode of reservoir area is on the order of 1–10 km2 in GRanD but was refined by 1 order of magnitude to 0.1–1 km2 in GeoDAR.

If we group the global dams by their documented main purpose, we observe in Fig. 15 that GeoDAR improved GRanD unanimously in both dam count and storage capacity for all main purposes (Fig. 15). For the same reason as explained above (i.e., the added reservoirs are small), the increases in dam count appear more prominent than those in storage capacity, and the increases in storage capacity from GRanD to GeoDAR are overall more evident than those from GeoDAR to the ICOLD WRD. The exception is the dams with “others” or “unknown” purposes, whose total storage capacity in GeoDAR is lower. This is because when GRanD and WRD records conflict with each other in the GeoDAR harmonization process, the attribute values in GRanD took precedence only if they are available or valid (“others” or “unknown” was considered to be an invalid reservoir purpose). Assuming that reservoir operations vary by purpose, this unanimous improvement of the spatial inventory for all reservoir purposes, in conjunction with satellite-observed water budget variations, can help us better generalize reservoir operation rules which are critical to improving water management.

The recently published GOODD (V1) dataset (Mulligan et al., 2020) includes 38 667 dam points in the world, which were digitized by scanning through Google Earth imagery with support of regional inventories and the Shuttle Radar Topography Mission Water Body Dataset (SWBD; Farr et al., 2007). Despite lacking essential attributes, GOODD is thus far the most comprehensive global inventory of dam locations and catchments. The digitization was performed during 2007 to 2011 and was later updated in 2016. This means that reservoirs postdating 2016 were not yet included in the dataset. The completeness and accuracy of GOODD also depend on the sizes of the dams or reservoirs. According to Mulligan et al. (2020), the resolution and quality of available Google Earth imagery during the digitization period were low in some parts of the world (such as China), and an experiment in the US showed that detectable dams and reservoirs from low-resolution imagery (e.g., Landsat Geocover 2000) may require a reservoir length greater than 500 m and a dam width greater than 150 m. These minimum-size criteria do not necessarily duplicate those of the ICOLD WRD, which instead emphasize the reservoir storage capacity and dam height (see Sect. 1).

Because of these digitizing limitations and criterion difference, the dam points in GeoDAR are spatially complementary to, rather than always duplicated by, those in GOODD across many regions. Figure 16 identified four examples in Cerrado Brazil, northern China, southwestern France, and northern Pakistan, where a large proportion of the GeoDAR dams were not digitized by GOODD. Some of the dams that only appear in GeoDAR also comply with the minimum size criteria of GOODD, and examples are those enlarged in the right panels, except the Duber Khwar Dam in Pakistan (35.119∘ N, 72.927∘ E; Fig. 16j), which was completed more recently in 2014. Since the area of the Duber Khwar Reservoir (about 0.05 km2) is smaller than the resolution of HydroLAKES (0.1 km2), and the dam completion year overlaps with the image acquisition period of the UCLA Circa 2015 Lake Inventory (from May 2013 to August 2015; Sheng et al., 2016), GeoDAR georeferenced the dam point but did not successfully retrieve the reservoir polygon.

To approximate how GeoDAR and GOODD complement each other globally, we intersected both dam datasets with the 30 m resolution UCLA Circa 2015 Lake Inventory. As a result of manual snapping to the 30 arcsec HydroSHEDS streamflow network (Lehner et al., 2008), some of the points in GOODD ended up having substantial geographic offsets from the actual dam or reservoir locations. For a pilot experiment, we applied a 1 km tolerance (about 30 arcsec on the Equator) when intersecting the UCLA Circa 2015 Lake Inventory with GOODD and kept a 500 m tolerance as used in Sect. 2.5 for intersecting the lake inventory with GeoDAR. The result shows that among the 55 000 or so water bodies that intersect either dataset, 80 % intersect with GOODD and the other 20 % with GeoDAR alone. These statistics imply that GeoDAR may have an ability to expand the number of dams in GOODD by roughly 25 % (i.e., 20 % divided by 80 %). Since we applied a larger tolerance for GOODD, this estimated expansion by GeoDAR is likely conservative (considering that the number of GOODD-intersecting reservoirs may be overestimated). If a 500 m tolerance is used for both intersections, the expansion by GeoDAR will increase to roughly 45 %. In addition to the expanded spatial coverage, GeoDAR indexed each georeferenced dam point to a WRD and/or GRanD record and thus enabled access to multiple attributes, whereas GOODD carries no attribute information except the delineated reservoir catchments. These regional and global comparisons suggest that, even just with the geometric dam points, GeoDAR is not a simple replication of GOODD but instead complements GOODD for an improved spatial coverage and density of global dams.