| 08 Aug 2025
Decade-long isotope dataset of rainfall and non-rainfall waters in the central Namib Desert
Abstract. Drylands are essential Earth System components, and dryland dynamics are strongly controlled by water availability. Long-term ground observations of hydrological parameters are often lacking in most dryland ecosystems. A series of foundational and unique water input data, including both rainfall and non-rainfall (i.e., fog and dew) components, in hyper-arid desert environments have been presented in this database. These observations provide comprehensive isotope measurements of rainfall and non-rainfall water resources, which are essential to evaluate the impacts of extreme climate events on the water cycle. The database comprised three key components: (1) a decade-long (2014−2023) event-based stable isotopes (δ2H, δ18O, and d-excess) in rainfall and non-rainfall waters at Gobabeb, (2) a two-month spatial isotope dataset (δ2H, δ18O, and d-excess) of fog collected from the central Namib Desert in 2016 and 2017, and (3) a six-decade (1963−2023) temporal data documenting monthly fog and rainfall amounts as well as annual Kuiseb River flooding events at Gobabeb. The detailed stable isotope data included 585 fog samples, 71 rainfall samples, 115 dew samples, and 13 groundwater samples over the past ten years (2014−2023) in the Namib Desert. Detailed descriptions of the study sites, sampling procedures, analytical methods, and data quality control are provided in this study. The uniqueness of our long-term dataset makes it an important resource for future studies investigating hydrological processes in drylands and their responses to climate change. The DOIs of the dataset can be obtained in the “Data availability” section.
Review of «Decade-long isotope dataset of rainfall and non-rainfall waters in the central Namib Desert» by Li et al.
This paper presents a database consisting of three main components:
(1) a decade-long (2014−2023) event-based stable isotope record of rainfall and non-rainfall waters at Gobabeb,
(2) a two-month spatial isotope dataset of fog collected in the central Namib Desert in 2016 and 2017, and
(3) a six-decade (1963−2023) temporal dataset documenting monthly fog and rainfall amounts as well as annual Kuiseb River flooding events at Gobabeb.
The isotope dataset itself has been published on Pangaea (Li et al. 2025b, hereafter Li25b), and the present paper can be viewed as an extended description of that dataset. There is, however, substantial overlap with Li et al. (2025a, hereafter Li25a), in which components (1)–(3) are already described in detail. The only substantive addition in the current paper is the inclusion of data from 2022 and 2023. While the publication of this isotope dataset is certainly valuable, several important issues require attention.
First, the dataset does not include the fog categories (radiation, mixed, advection) that the authors have assigned to individual events. Since these categories are derived from the isotope data itself, and their meteorological basis is, in some cases, questionable, they should be made available alongside the dataset. For example, the event on 16.12.2014 is classified as radiation fog, despite being associated with a wind speed of 7 m/s (Kaseke et al. 2017). Making these classifications transparent is essential so that their validity can be independently evaluated.
My principal concern relates to the description of dataset (1). In Li25a, the authors state with regard to the isotope data collected at Gobabeb: “Additionally, we opportunistically recorded event‐based fog amount data using the 1 m² SFC as a backup between October 2014 and December 2022.” Similarly, Li25b reports that the temporal fog samples at Gobabeb were collected using a standard fog collector (SFC). However, it now appears that from 2020 onwards, the samples were not obtained with an SFC, but rather from rooftop drippings collected by one of the co-authors. The number of fog days based on these rooftop samples is much higher than that derived from fog precipitation measured with the Juvik collector. Moreover, there are several cases where fog samples are reported on days when meteorological data at Gobabeb show no indication of fog (e.g. in June 2023, 6 out of 15 reported events lack supporting evidence from fog precipitation, visibility, cloud base height, air temperature cooling rate, or longwave downward radiation).
There is no reason to doubt that water was indeed dripping from the roof, and these observations may point to interesting scientific questions—for example, whether they represent dewfall triggered by a nearby fog front, or processes related to cooling or changes in specific humidity. Nevertheless, omitting this information about the sampling method is problematic, as it affects the interpretation and comparability of the dataset.
Finally, the explanation given in Li25a—that the higher number of fog days is due to the greater sensitivity of the SFC to low-intensity southerly fog compared to the Juvik collector—appears inconsistent with the actual sampling methods employed after 2020. This discrepancy is confusing and, in its current form, undermines confidence in the dataset description.
Kaseke K. F., Wang L. X. and Seely M. K. (2017): Nonrainfall water origins and formation mechanisms, Science Advances, 3(3) (doi: 10.1126/sciadv.1603131).
Li Y., Wang L., Diersing C., Qiao N., Yi L. Maggs-Kölling G., Marais E. (2025a): El Niño intensified fog formation in the Namib Desert, Earth's Future (doi.org/10.1029/2024EF005725).
Li, Y., Marais, E., Wang, L. (2025b): Decade-long isotope dataset (δ2H and δ18O) of rainfall and non-rainfall water samples in the central Namib Desert, Namibia [dataset]. PANGAEA, (doi.pangaea.de/10.1594/PANGAEA.981180).