The manuscript presents a new ice breakup dataset for the Kokemaki river and compare it with dates of ice breakup at two other rivers - Aura and Torne, as well as with air temperature values. The long time series (220-320 years) constructed from archives are very helpful to get an assessment of long-term variability of ice breakup.

General remarks. 

The manuscript often gets very descriptive, especially starting from section 4 (Results), and one may easily feel overwhelmed by various details and lose track of what is important and what is just a side observation. I strongly suggest to streamline the manuscript by a) shortening some parts and focusing only on important issues, b) clarifying importance of some facts (and not just enumerating them) and c) providing intermediate conclusions. 

Dates of ice breakup are dependent on various factors, including earlier or late ice formation, severity of winter (reflected in ice thickness). Influence of air temperature on ice growth and decay may further be modulated by snow depth (as you mention in the "Discussion"). I understand that the authors do not have the data on ice formation dates and other parameters to take all these factors into account. However it would be very helpful to indicate - even roughly - the duration of presence of ice cover on rivers (or average dates of ice formation), as well as maximal ice thickness, typical range of depth of snow on ice. Otherwise ice breakup dates are somewhat taken out of context. 

I suggest to provide an overview map with location of the three rivers, otherwise it is not easy to imagine their location. Provide at some point in the introduction some rough average values of river discharge, river length and size of the watershed for each of the three rivers. 

Also - if you have those data - provide in sub-sections 2.1-2.2-2.3 some estimates of river depth at place of measurement/observations.

Climatic correlations - why use air temperature for fixed months for correlation and not do some dynamic approach (if ice breakup this winter happens in month X, then we will look at air temperatures in months X and X-1)? Or - better still - do some integration/aggregation of air temperatures, such as sums of negative/positive degree-months or similar, to account the severity of each winter and estimate influence of spring air temperatures on ice breakup. 

Specific comments.

Line 9 - please specify that Pori it is a name of settlement and not a person's name.

Abstract - provide at least rough location of the three rivers, distance between them.

Line 47-48  please clarify the difference between "ice breakup series" and "observations"

Line 53 - please clarify. Ok, it did not escape the boom, but then what - the time series are not homogeneous? Or else?

Line 68 - "because of their length" - rephrase so that it is clear that it is length of time series and not length or rivers

Line 101 - put comma after "by temperature"

Line 120 - here and for other sub-sections I would suggest to put river name first and then city name (and also specify that it is a city/settlement)

Line 141 - specify that it is 80 km upstream. Any significant tributaries between power plant and Torne?

Line 141 - observation site - Torne or Tornio?

Line 160-161. Rephrase this sentence ("For example…) or split in two sentences..

Lines 169 and 170 - so this river has a delta or estuary?

Line 283 - please remind the reader on which river the power plant is located

Line 316 - "16 respectively" => "16 and respectively"?

Line 318 - please remind what are these distances

Lines 360-370 (and also in general). Please use expressions such as "worth noting", "noticeable", "remarkable", "noteworthy", "notable" etc sparingly. See also my general remark on the descriptive style of the manuscript.

Lines 419 and 420. Temperature and breakup dates are presented in Fig 4 but they are not discussed in the text. Is figure 4 really necessary then?

Figure 1 - upper panel, and also Figure 5: Tornio or Torne river?

Overall this paper presents a new long-term river ice record and examines changes in the breakup timing between three northern rivers.  It first places the records in historical context and then compares them to temperature changes, and examines temporal changes including extreme timing of ice off events.  Overall, while interesting, the manuscript reads as a bit verbose, and qualitative in places where perhaps some quantitative analysis could be used. There is some interesting information presented, and the long-term datasets analyzed together are a new contribution to the literature, but I think some of the qualitative description could be reduced to shorten the read.

The main thing I found to be lacking is the treatment of precipitation. The authors indicate quite correctly snow is important for breakup timing, not just with respect to delaying melt through high albedo as mentioned but also for the effect on the ice thickness. Also, when discussing ice break up – do the extreme events link to precipitation (rain or snow?) as well as temperature? Only temperature is compared. Can you include snow changes as well? derive a simple metric, perhaps total winter snowfall from typical freeze-up to break-up months? Or total winter/spring precipitation to include spring rainfall? I imagine the records may not go back as far as the entire river ice record, perhaps that is why they are not included. But some inclusion of precipitation in the analysis is important and I feel that is should be addressed with more detail when revised. 

The authors should add a map to show where the rivers are to provide spatial context.

Some details are missing from the methods that I think will add more clarity. 

Adjusting the dates to the vernal equinox is interesting and could use a bit more explanation of how that is done.  When looking at the tables, its clearer that is the difference between identified date and the equinox is used, but 3.2 as written is not very clear. How was the actual equinox determined? Is this a dataset available somewhere with the timing listed for the relevant years before the calendar change?   As someone who has never worked with datasets going back that far historically, I found this very interesting. 

The temporally extreme events – I found this section confusing and after several reqds still do not understand why 2 of the list of years were used. Can this be re-written more clearly? or explained why only 2 were used? 

“In this analysis, we used the calendric dates to rank the breakups”. – does that not skew some of the early records?

Section 3.5 – what is the model used for 1960-2020? More information is needed here to describe the data fully.  

Specific comments:

78-80: Are the other two study rivers regulated? I think not based on the sentence about the power plant boom but confirming would be good in the text.  Comments later on about how the power plant may have changed the timing of break up but they focus on the thermal effects – did regulation have any effect with respect to water level? 

112-113: “In Aura River, the records suggest that thermal breakups have delayed the ice-off date and this is because spring is the driest season in Finland (Irannezhad et al., 2014).”  This could use a small clarifying sentence added that thermal break up would be later than mechanical break up since its thermodynamic rather than dynamic and the dry spring would reduce the runoff/melt. Its more or less stated in the sentences earlier, but an explicit sentence stating that would be useful. Also why more thermal now? is spring becoming drier?

155-118: This whole bit is unclear to me.  Lack of clear breakup dates because thermal melt made it challenging to determine the timing in Pori. Thermal break ups are delaying ice off because its dry. Then ‘this’ is because you are comparing ice off to ice breakup? I think you need to add some more info here on the timing difference between ice off and break up – you are comparing 2 different things. how much time generally passes between the two events on the Aura river? Is it consistent? (see comment later on this as well).

255: How did you count thermal break ups and distinguish from dynamic?  are they distinguished in the records?  

316: April 16 and 15 respectively?

343-346: Can you really compare the moving stake to breakup? is there a consistent offset?

402-404, 433:  How do you define strong? Common pattern of variability, can you analyze statistically to quantify this?

449-450 and onwards: why is the change in the Kokemaki River actually an overestimation of climate change? Explain?  

476-479: Talking about projected temperatures. How about projected precipitation and possible effects on breakup?

516: “Arguably, the warmer climate that is dominating in the south is changing more rapidly, and with less predictability, than the colder climate dominating in the north. A similar latitudinal shift has been noticed in Swedish lakes (Hallerbäck et al., 2021; Weyhenmeyer et al., 2005). “  Is this not because the temperature is closer to 0 in the lower latitude river reaches so a small temperature shift will have a more pronounced effect on the ice? This is seen in lake ice in near-zero regions compared to northern regions. 

Table 2 has lines every 10 records but table 2 does not, I would suggest removing them from Table 1.

The authors present a new ice record extending back several centuries and put this new time series into context with two other rivers in the region with time series extending back centuries. It is clearly a lot of effort to assemble, revise, and validate a new record. The manuscript is quite descriptive and verbose, and could benefit from more formal quantitative analysis to support the data. The manuscript could also use better contextualization for readers outside of Finland who are not familiar with the geography and history to better follow the manuscript. Nonetheless, it is an interesting manuscript and would be a good addition to the literature.

On line 36, the authors state that the “databases are not updated with observations from the first two decades of the 21st century.” This is no longer true as the database for the National Snow and Ice Data Centre have been updated in 2020 and there have been a variety of publications using ice databases that have provided their data online from the past 2 decades. Please update this sentence to reflect updated databases and recent publications.

The new time-series at Kokemäki River is amazing! Please provide some historical context of this ice time series. Who collected this time series? How was it discovered? What kinds of observations were made? Where were the observations taken? Was the same methodology used throughout the time series? How long does the breakup process take for the river?

Since one of the goals was to verify this new time-series, please provide further qualitative and quantitative evidence of how the data were verified, how closely they matched between different newspapers/records? I was also curious when there were multiple sources of ice-off, were the dates the same? Also, was the same definition used by each source over time? Did they examine breakup date at the same location and in the same way?

On line 107, thermal breakups were introduced. Please provide more specific details related to the thermal inputs to the river. What thermal breakups are relevant in Kokemaki River and why? This seems interesting, but much too vague for readers not familiar with the detailed geography and history of Finland to follow.

On line 234, could you please clarify what is meant by previously published ice breakup dates and which river is this referring to?

In the methods, it has become clear that there were changes in the location and observations for ice breakup over the years? How were the changes in location and definitions reflected in the dates and patterns of ice-off? Did they coincide with extreme events or breakpoints in the data? And how else could they have impacted the uncertainty around ice-off dates?

I liked the introduction of extreme events, but I was hoping to see a more quantitative analysis than simply identifying the 30 most extreme years. Perhaps more formal analysis could be done here to quantify if there are more extreme years in certain decades/periods than others or expected by chance. Also have the number of extreme events increased over time? Further quantitative analysis would be appreciated here.

On line 283, the analysis with the hydroelectric plant was introduced. Is this power plant only relevant to Kokemaki? It doesn’t seem so as earlier it was stated that there are 4 power plants on this river. Are there any other power plants in the other rivers? Why was only 1 power plant included in this analysis and not the others?

I was curious how far these rivers and sites are from one another? If the sites are more than a grid cell apart, I was curious why air temperatures were used from the same station.

Please provide a map of these rivers and also include some of the features that were discussed in the manuscript.

The impact of the power plant is much too descriptive. Please provide some more formal quantitative analysis to illustrate whether findings were significant. Earlier in the manuscript, many other factors that have contributed to land use changes and warming were discussed, but not included in the analysis, including urbanization, land use change, and climate change. How have these factors contributed to the ice-off dates?