The study aims exploring and quantifying the degree of error between researchers assigning ordinal-scale indices to a historical documentary dataset. Two teams of raters were asked to produce a five-category annual rainfall index series for a dataset consisting of transcribed narrative descriptions of meteorological variability for 11 rain-years’ in nineteenth-century Lesotho. The authors conclude that variability between researchers should be considered minimal where index-based climate reconstructions are generated by trained historical climatologists working in groups of two or more.

The study should be accepted with small changes indicated below:

The different preconditions in the derivation of temperature and precipitation indices should be worked out more clearly:

Lines 49-56 

“This approach has been adapted for regions with less rich documentary evidence”, “The Pfister is approach is mainly tailored to reconstructing temperatures for regions with rich documentary evidence and long series of homogenized instrumental measurements (e.g. Pfister, Wanner 2021). In such cases proxy information often allows estimating temperatures for specific months or seasons by using the calibration verification approach (e.g. Dobrovolný 2010). In such cases, the potential bias in classification is very small for trained historical climatologists, as the narrative record and the proxy need to be consistent and meteorologically meaningful. This approach has been adapted for regions with less rich documentary evidence, or a seasonal skew to the available climate descriptions, through a reduction in the number of index categories (e.g. to five or three classes) and/or the temporal resolution of the reconstruction (to seasonal or annual).

The situation is different for classifying precipitation. Proxy-data such as information on floods and droughts or the number of rain-days may hardly be calibrated, as precipitation is rather small-scaled in comparison to temperatures and because long homogenized instrumental series of precipitation are quasi non-existent. The study by Dobrovolný et al.(2015), which is perhaps the most sophisticated approach of this kind in Europe, only indicated acceptable reconstruction skill for seasonal precipitation indices in JJA and annual values.

Response: Thank you for this observation. This distinction is important but not central to the main aim of the paper. Accordingly, we have added some brief text to the middle and end of para 2 in section 1, as follows: 

(middle) “The Pfister method is mainly tailored to reconstructing temperature variability for regions with rich documentary evidence and long series of instrumental data (Pfister and Wanner, 2021). Central to the method…”

(end) “The reconstruction of precipitation is more challenging using the Pfister method, since (i) rainfall often varies over smaller spatial scales than temperature, (ii) proxy data such as drought or flood magnitudes are less easy to calibrate, and (iii) the long instrumental series required for calibration are less common than those for temperature. The study by Dobrovolný et al. (2015) of precipitation variability over the last 500 years in central Europe, for example, only identified an acceptable level of reconstruction skill for seasonal precipitation indices in JJA and for annual precipitation values.”

It should be worked out more clearly, also in the abstract, why estimates of historical precipitation conditions in a country of the Global South are significant for the present situation.

Response: We have added the phrase “…and effectively extend the instrumental record” to the opening sentence of the abstract. We have inserted the following sentence at the end of para 1 in section 1: “The reconstruction of climate indices is a useful tool for examining climate variability during the pre-instrumental period, and is particularly valuable for regions, including many in the Global South, where lengthy meteorological records are lacking.”

    

A map should be included showing the location of Lesotho in southern Africa.

Response: We have inserted a map as Figure 1 and added a caption to the manuscript.

An example of a source illustrating the nature of the narratives should be included.

Response: We have inserted an image of an example source as Figure 2 and added a caption to the manuscript.

Suggestion for small changes:

Line 45: “Under the Pfister method, indices are normally” might be replaced by “The Pfister Indices, as Mauelshagen (2010) named them”, are normally generated,

Response: Thanks for this suggestion. This sentence now reads: “Pfister Indices, as named by Mauelshagen (2010), are normally generated…”

Lines 49-51: ”relevant phenomena (e.g. the timing and duration of snowfall, or various plant-phenological indicators) “might be replaced by . “regionally relevant proxy data (e.g. plant-phenological observations, the duration of snow-cover and the freezing of water bodies)…

Response: This sentence has been amended as suggested.

Additional references:

Mauelshagen, F. (2010), Klimageschichte der Neuzeit, 1500–1900. Darmstadt: Wissenschaftliche. Buchgesellschaft, 2010

Pfister C. and Wanner H. (2021). Climate and Society in Europe

Response: These have been added to the manuscript.