Dataset of cropland cover from 1690 to 2015 in Scandinavia

High-resolution historical land cover datasets are essential not only for simulations of climate 15 and environmental dynamics, but also for projections of future land use, food security, climate and biodiversity. However, widely used global datasets are developed for continental-to-global scale analysis and simulations and the accuracy of global datasets depends on the verification of more regional reconstruction results. In this study, based on the collected statistics of cropland area of each administrative unit (Parish/ Municipality/ County/ 20 Province) in Scandinavia from 1690 to 2015, the cropland area at the administrative unit level was allocated into 30-arc second grid cells. The results indicated that the cropland area increased from 1.81 million ha in 1690 to 7.10 million ha in 1950, then decreased to 6.02 million ha in 2015. Before 1810, cropland cover expanded in southern Scandinavia and remained stable in northern. From 1810 to 1910, northern Scandinavia experienced slight 25 expansion and the cropland area increased rapidly in the southern part of the study area. Then, cropland area changed gently. After 1950, cropland area began to decrease in most regions, especially in the east of Scandinavia. When comparing HYDE3.2 with this study, differences https://doi.org/10.5194/essd-2020-187 O pe n A cc es s Earth System Science Data D icu ssio n s Preprint. Discussion started: 6 August 2020 c © Author(s) 2020. CC BY 4.0 License.


5
Anthropocene was defined as a new epoch of geologic time, partly because human influenced land is a major component of anthropogenic global changes in the earth system, since 2000 (Lewis and Maslin, 2015;Crutzen and Stoermer, 2000;Verburg et al., 2016). During AD 800 ~ AD 1700, according to Pongratz et al. (2008), 5% of the area covered by natural vegetation brought under human land use, compared to 44% in the following 300 years. The decrease of 10 natural vegetation is accompanied by an increase in cropland area. From AD 1000 to AD 1700, global cropland grew slightly from 1% to 2.3% of the global land area, then it occupied 4.5% and 11% in AD 1850 and AD 2015, respectively (Klein Goldewijk et al., 2017).
For example, land use change likely led to an increase of the Earth albedo with a radiative forcing of -0.15 ± 0.10 W m -2 since 1850 (Myhre et al., 2013). Over the years 1980-2005, changes from mid-latitude natural forests to cropland and pastures were accompanied by a 20 reduction of latent heat flux and an increase of sensible heat flux in summertime (Findell et al., 2017). The atmospheric CO2 increase above pre-industrial levels was mainly caused by the release of carbon from ALCC and net emissions from land cover change were greater than 50% of the total before 1940 (Ciais et al., 2013;Le Qué ré et al., 2018;Houghton, 2018). The above conclusions of climate and environmental dynamics were all made by using 25 high-resolution historical land cover datasets. Historical land cover change information is also essential as baseline analysis for projections of future land use, food security, climate and biodiversity (Hurtt et al., 2011;Foley et al., 2011;Ellis et al., 2012;Brovkin et al., 2013;Mehrabi et al., 2018;Fuchs et al., 2015). Precise information on land cover is required for https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License.
running earth system models.
Satellite data can provide information on changes in land cover with high-resolution, however, it is impossible to map land cover earlier than the mid-1970s (Liu and Tian, 2010;Roy et al., 2014;Moulds et al., 2018). Therefore, satellite data was combined with historical statistics and inventory data in order to produce spatially explicit land cover datasets which cover 5 longer time periods. Using combined sources and hindcasting methods, Center for Sustainability and the Global Environment (SAGE) (Ramankutty and Foley, 1999) and History Database of the Global Environment (HYDE) (Klein Goldewijk, 2001) was produced as two representative datasets of global land use/cover. SAGE covers the period between 1700 and 1992 with a 0.5°×0.5° resolution, including cropland, pasture and forestland 10 (Ramankutty and Foley, 2010). The latest version of HYDE (HYDE 3.2) has a 5′×5′ resolution, covers the period 10000BC -AD2015 and consists of cropland, pasture and built-up land (Klein Goldewijk et al., 2017). Based on SAGE and historical population data, PJ dataset was produced, which covers the period AD 800 -AD 1700 (Pongratz et al., 2008).
However, global datasets were developed for continental-to-global scale studies (Klein Goldewijk, 2001;Ramankutty and Foley, 2010;Pongratz et al., 2009a). Assuming near-constant cropland per capita and using population data to estimate historical land use induced large uncertainties and limitations in presenting the details at the regional scale 25 (Klein Goldewijk and Verburg, 2013). Many regional land use reconstructions illustrated that global datasets had unneglectable discrepancies in reflecting the spatial patterns of land use at the regional scale over history, especially for cropland. Historical document-based reconstructions concluded that SAGE, HYDE, and PJ had drawbacks to capture the spatial https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License. distribution of the historical cropland change in China (Li et al., 2010;Zhang et al., 2013;Li et al., 2015;Wei et al., 2019;Li et al., 2019). In the US, the HYDE maps greatly underestimate crop density in high cropland coverage regions but overestimate it in low-density regions during 1850-2016 (Yu and Lu, 2018). Neither KK10 nor HYDE captures the fine-scale spatial pattern of open land as inferred from the pollen-based land cover 5 reconstructions in Europe for four preindustrial time windows (Kaplan et al., 2017).
Uncertainties in global datasets could be transferred into higher uncertainties in quantifying climate and environmental effects of ALCC at both local and regional scales (Yang et al., 2018;Lejeune et al., 2018;Yu et al., 2019). Therefore, the PAGES LandCover6k and related projects aim to improve ALCC history at both regional and global scales based on empirical 10 data (Gaillard et al., 2015a;Widgren, 2018a). Errors can be assessed or corrected by using the regional quantitative reconstructed land cover data and regional agrarian history maps (Widgren, 2018b;Fang et al., 2020).
In Scandinavia, the pollen-based reconstructions indicated that the cover percentage of cereals was very high in southern Sweden and Denmark (Nielsen et al., 2012;Gaillard et al., 2015b). 15 The comparison of the pollen-based reconstructions with historical map-based estimates of land use history in the South Swedish Uplands over the past 200 years showed that the pollen-based estimates overestimated the cropland and grassland cover in some sites (Mazier et al., 2015). To facilitate simulation studies with high-precision regional input data, the dataset of cropland cover change from 1875 to 1999 in Sweden and Norway was developed 20 (Li et al., 2013). Materials from mainly two sources were collected, including the data of Sweden from the official agricultural statistics and the data from Norwegian Social Science Data Services (NSD). To access orderly time intervals, 6 time slices were chosen. Methods of seed-cropland conversion, data interpolation and allocation, and data gridding were used to produce the cropland dataset on the spatial resolution of 0.5 degrees. 25 The main objective of this study is to provide a cropland cover dataset with high precision and spatial resolution in Scandinavia (includes Sweden, Norway and Denmark) over the past 300 years. In this paper, on the basis of the cropland datasets in Scandinavian Peninsula from 1875 to 1999 (Li et al., 2013), we extended the period to 1690-2015 and added Denmark to https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License. the study area through validating and integrating data from multi-sources. The spatial resolution of our dataset was improved to 30-arc second. Our newly developed dataset of cropland cover would give more detailed information on temporal and spatial patterns of cropland change in Scandinavia.

Data sources
5 Data sources in this study are shown in Table 1. Besides the cropland data of the Scandinavian Peninsula (Sweden and Norway) after 1875 processed by Li et al. (Li et al., 2013), the cropland data were from different sources. For Sweden, all the data before 1875 were from the Svensk Nationell Datatjänst (Swedish National Data Service, termed SND, https://snd.gu.se/en/catalogue/study/SND0910). Based on tax records, historical maps, land 10 survey records and inventories of farmers, Palm et al. (2014) developed agricultural statistics (The database Sweden 1570-1810: population, agriculture, land ownership) covering all parishes within Sweden's contemporary boundaries and the periods between 1570 and 1810.
In the database, cropland was called "åker" in Swedish, which was also used in the data sources from 1875 to 1999 in Sweden. However, the "total å ker" in the dataset recorded the 15 volume of seed used in each parish (socken) but not the cropland area. The administrative division map of 1750 was used as the base map for cropland data before 1875.
For Norway, cropland data in 1665 and 1723 was from Statistiske studier over folkemaengde og jordbrug i Norges landdistrikter i det syttende og attende aarhundrede (Statistical studies on population and agriculture: in the rural areas of Norway in the seventeenth and eighteenth 20 centuries, Aschehoug, 1890). Cropland data in 1809 was from the study of Hovland (Hovland, 1978). In the above two sources, the volumes of different types of seeds, such as wheat, rye, barley, oat, peas and potatoes were recorded but not the cropland area. The cropland data of 17 counties (amts) were presented. The administrative map of 1875 which was from Norwegian Centre for Research Data (termed NSD, https://nsd.no/nsd/english/) was used for 25 cropland data before 1875. Rydé n Rømer, Aalborg, Bernd Münier and Morten Stenak, Roskilde (Odgaard and Rømer, 2009). In detail, data in 1800 was from the map Videnskabernes selskabs kort (VSK) in 1762-1806 and further developed from agricultural statistics. Data in 1881 was generated from the national statistics. The data in 1998 was merged from maps and statistics. The data in 1800, 1881 and 1998 had a smallest spatial unit of "sogn" (parish). The cropland was called 10 "Agerjord" and it can be divided into two subgroups, namely "Besået areal" (the areas annually sown with various one-year or two-year crops) and "Graes i omdrift" (the areas for shorter or longer time usually from 1 to 4-6 years which is for grazing or in various types of fallow before land again plowed up for sowing). "Besået areal" and "Graes i omdrift" corresponds to the cultivated land and the fallowed cropland in 1688, respectively. The 15 cropland area data in 1907 was from Statistisk Aarbog 1912 (Statistical Yearbook 1912, Danmarks Statistik, 1912 and the spatial resolution is "amt" (County). The cropland area data in 1936, 1950 and 1980 were from agricultural statistics of Statistiske Meddelelser (Danmarks Statistik, 1936, 1950and 1980. The spatial resolution of cropland data in 1936, 1950 and 1980 was "amt" (County), "amtsrådskreds" (County council) and "kommuner" (Municipality), 20 respectively. The base maps were downloaded from the HisKIS network Statistics Denmark (https://www.dst.dk/) and the cropland area of each province (Landsdele) was recorded. Base maps were extracted from the GADM database (www.gadm.org)  1936, 1950and 1980Denmark 1936, 1950, 1980County council;Municipalit y Danmarks Statistik, 1936, 1950and 1980 Cropland in Scandinavian Peninsula Sweden, Norway 1875, 1910, 1930, 1950, 1980, 1999

Data preprocessing
The main goal of the data preprocessing is to map the cropland area change at the parish/county level from 1690 to 2015 in Scandinavia based on the collected data sources and 5 base maps.
According to the years when cropland area data are available in Sweden, Norway and Denmark, we chose 10 points of time that could present the trend of cropland change from 1690 to 2015. Compared to the data sources in Norway and Denmark, the numbers of cropland data from the data sources of Sweden were the most abundant and complete, so we 10 selected 10 points of time based on the data sources in Sweden, which were 1690, 1750, 1810, 1875, 1910, 1930, 1950, 1980, 1999 and 2015. 1690, 1750 and 1810 corresponds to 1665, 1723 and 1809 from the data sources in Norway. 1690, 1810, 1910, 1930, 1950, 1980, 1999 and 2015 corresponds to 1688, 1800, 1881, 1912, 1936, 1950, 1998 and 2016 from the data sources in Denmark. 15 Because the Swedish data in 1690, 1750 and 1810 from SND only provided the amount of seeds sowed into the land, we used the formula of 1 barrel of seed = 4936m 2 of cropland = 0.4936 ha of cropland (Cardarelli, 2003) to obtain the cropland area of each parish. The Norwegian data in 1665, 1723 and 1809 also showed the volume of seed but not the cropland area. We collected the relationships between the volume of seed and the cropland area from 20 four sources (see Table 2). The values of liter per maal (1 maal=10 hectare) regarding seven https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License. types of seeds were close to each other, except for the value based on the statistics from NSD.
We chose the values of liter per maal in 1835 from Statistiske Oversigter 1914 (Aschehoug, 1914) because 1835 was the closest year to the 3 points of time, 1665, 1723 and 1809. For the data in 1875, we used the values of liter per maal in 1865 to convert the volume of seed to cropland area. For data sources in the remaining years and in Denmark, cropland areas were 5 recorded and we unified the units to million hectare (million ha). There was no record for cropland area of Denmark in 1750, we therefore assumed that the cropland area change rate from 1690 to 1810 was constant and computed the cropland area in 1750.
Table2 The relation of maal to liter from different sources in Norway 10

Sources
Oversigt over det norske landbruks utvikling siden 1750 (Klokk, 1920) Norges Landbrug i Dette Aarhundrede (Smitt, 1888) Statistiske oversigter 1914 (Aschehoug, 1914) NSD Liter per maal\Year 1896-1900 1901-1905 1907 1866-1875 1835 1865 1907 Hvete ( To map the spatial patterns of cropland distribution in Sweden, administrative map at the found their corresponding cropland area data. Because the cropland data before 1875 was at the county (amt) level and the administrative division didn't change dramatically from 1690 to 1875 in Norway, administrative map in 1875 from NSD's kommunedatabase was used for data in 1690-1810 as the base map.
For Denmark, cropland area data in 184 "ejerlavs" was missing in 1688. The missing data 10 were interpolated based on the cropland fractions of their neighboring "ejerlavs" in 1688 and the cropland changes from 1688 to 1800. Using the same method, 56 missing data in 1800 and two missing data in 1881 were also interpolated. We selected administrative map in 1688 from the HisKIS network as the base map for cropland data that year. Cropland area data in 1800, 1881 and 1999 had their corresponding base maps (Odgaard and Rømer, 2009). The 15 base maps for data in 1936, 1950 and 1980 were from the web of Danish Geodata Agency (https://eng.gst.dk/).

Cropland area allocation
As the cropland area data of each administrative unit cannot be used as input for the climate 20 and environment simulations directly, we allocated the cropland area into 30 arc-second (~1km) grid cells based on the cropland area allocation model built by Li et al. (2015).
In the allocation of cropland, we have allocated all historical cropland within the maximum extent of cropland in modern times. This is a source of error since at the maximum extent of cropland in historical times more marginal lands may have been cultivated (see more under 25 Discussion). However, at the scale in which we present the data the error is probably negligible. The 300m CCI-LC maps (http://maps.elie.ucl.ac.be/CCI/viewer/index.php) https://doi.org/10.5194/essd-2020-187 We analyzed the factors affecting the distribution of cropland. In previous studies, elevation, slope, climate, soil, water and population was used as causes that relate to the change of the 15 spatial distribution of cropland (Klein Goldewijk et al., 2011;Li et al., 2015;Paudel et al., 2017). Agricultural land is usually constrained by both elevation and slope. People tend to be more inclined to start from areas with lower elevations and gentler slopes when cultivating land. Land with high elevations and slopes has negative characteristics that constrain cropland cultivation, which will only be used after low-elevation and gentle-slope land has been 20 cultivated. Elevation and slope also effects climates and water availability. Climate can be used as one of the secondary characteristics of elevation and slope that affect cropland distribution and it accounts for more of the regional differences in the types of crops grown.
Moreover, in Scandinavia, there is little climate difference in each administrative unit and we assumed the impact of climate for cropland distribution had been included in the effects of 25 elevation and slope in each administrative unit. Soil provide the physical base and nutrients for crops. Soil properties such as texture, fertility and organic-matter content impact suitability for growing crops, but it is not a limitation for land cultivation in most areas. With the development of agricultural technology, agriculture modification can make soils more https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License. favorable for growing crops. In addition, the soil in the cultivated area has been generally ripened. Thus, modern soil datasets are unsuitable to be used for cropland area allocation.
According to statistics, population in Scandinavia increased constantly from 17 th century to present (SCB, SSB and Statistics Denmark), while the cropland area decreased from 1950.
The continuous increase in population has inevitably led to the increase in urban land and the 5 reduction in cropland. Population growth was not the most important reason for cropland area increase after 1950, especially in Sweden and Denmark. However, the spatial distribution of population data before 1950 at a high spatial resolution is rather hard to obtain. Therefore, elevation, slope and water were selected as the factors in the cropland area allocation model . Used the NASA Shuttle Radar Topographic Mission (SRTM) 90m digital 10 elevation data (DEMs), we resampled the DEMs in Scandinavia to 30 arc-second resolution and normalized the value of elevation and slope using the following formulas: W(i) was defined as the water area in grid i, the value 0 indicates grid i is occupied by water, the value 1 indicates grid i is occupied by land.
The value of land cultivation suitability Suitbf1999(i), Suit1999(i) and Suit2015(i) of grid i before 1999, in 1999 and in 2015 were calculated using the following formulas, respectively : 20 1999 1999 1999 1999 For convenience, Suitbf1999(i), Suit1999(i) and Suit2015(i) were denoted by St(i) in the following formulas.
The total weight of each administrative unit for cropland area allocation was defined as 1, then the weight of grid i for cropland area allocation (wcrop(i)) and the cropland area of grid i (Cr (i)) became: 5 T(pn) is the total cropland area of administrative unit pn.
The cropland area of each administrative unit was allocated in 30 arc-second grids followed by the above formulas. However, in a few grids, the allocated cropland area was larger than the total area of grid i. Formulas of (3), (4), (5), (6), and (7) were used to allocate the over-allocated cropland area to the grids where the allocated cropland area was smaller than 10 the grids area. This step was repeated until the cropland area of each grid was smaller than the grid area. The cropland area allocation process was used for the 10 selected points of time and we created a cropland area dataset at 30 arc-resolution from 1690 to 2015 in Scandinavia.

Changes of the total cropland area in Scandinavia
The total cropland area change in Scandinavia during the period 1690 to 2015 is shown in

Sweden
In Sweden, about half of the area is covered by forest. Mountains, marshes and lakes together cover approximately one third. The cropland area accounted for 1.50%~8.13% of Sweden's 15 total land area over the past 300 years. In 1690 croplands were especially dense in south Sweden, especially around the lakes Vä nern, Vä ttern, Mä laren and Hjä lmaren and in Skå ne, reflective of the long history of cultivation. After that, the spatial patterns of cropland became more intensive in the southern Sweden and began to spread to the north. In 1750, we see the appearance of cropland in several grid cells of county of Vä sterbotten and Norrbotten and the 20 cropland fractions in these grid cells were above 10% since 1875. A large number of grid cells with more than 80% cropland were seen and increased in 1910, account for 55.65% of the total grid cells with cropland in Sweden. During 1910During -1950, spatial patterns of cropland distribution remained stable, except for slight increase in Vä sterbotten and Norrbotten counties in the north and gentle change in Skå ne (Malmöhus and Kristianstad counties) and 25 Gotland. In the following years, the cropland area declined in most regions and the percentage of grids cells with more than 80% cropland dropped to 20.17% in 1980 and 17.18%  Olsofjorden and Trondheimsfjorden started growing again and the dramatic growth was in the 15 southwestern of Rogaland. We also see the increased cropland in the coastal area with low elevation. The grid cells with more than 80% cropland accounted for 32.53% of the total grid cells with cropland in 2015.

Denmark
Denmark is among the most intensively cultivated countries in Europe. Long history of land 20 cultivation brought widespread cropland cover since 1690 in Denmark. Grid cells with more than 20% cropland accounted for 60.11% of the total number of grid cells and most high-fraction grid cells were distributed in eastern Denmark, since soil conditions were more rates of change of cropland areas in northern and eastern Denmark, but positive rates in western part, including Ribe, Ringkjøbing and Viborg Counties. But the changes of cropland areas of most grid cells were less than 20%. After that, increased afforestation and the need for areas for urban expansion and infrastructure had gradually reduced the agricultural area (Pedersen and Møllenberg, 2017). Although cropland areas in South Jutland, Ribe, 5 Ringkjøbing, Viborg, and South Jutland Counties increased moderately, those in other counties of Denmark declined from 1950 to 2015. We also saw the areas with more cropland had been changed from the eastern area with soil conditions suitable for cultivation to the western area dominated by sand.  1690-1810, 1810-1910, 1910-1950 and 1950-2015

Comparison of cropland change with HYDE3.2 5
As one of the most widely used datasets for global historical land use, HYDE has relatively high temporal and spatial resolution. Moreover, it keeps providing new versions with improved quality. The latest version is HYDE3.2, which covers the period from 10000BC to 2015AD and has the spatial resolution of 5'×5' (Klein Goldewijk et al., 2017). Thus, we compared our historical record-based reconstructed results with HYDE3.2 to further analyze 10 the accuracy of HYDE3.2. We selected 1700, 1750, 1800, 1880, 1910, 1930, 1950, 1980, 2000, and 2015  The total cropland data in HYDE3.2 were higher than those in this study before 1910 and the cropland area in HYDE3.2 was decreased but slightly increased in this study. The maximum absolute difference of the total cropland area was 0.91 million ha in 1880. However, we saw larger differences for Norway and Denmark between HYDE3.2 and this study. In Norway, the total cropland areas in HYDE3.2 were consistently higher than those in this study from 1700 10 to 1950. Especially in 1880, the total cropland area of HYDE3.2 stood at 0.77 million ha whilst the cropland area of this study was only 0.32 million ha. By contrast, the total cropland areas in Norway from HYDE3.2 were lower than those from this study. In Denmark, three stages were identified in HYDE3.2: rapid growth from 1700 to 1950, sharp decrease from showed that croplands were mainly distributed in southern Sweden, around Olsofjorden and Trondheimsfjorden in Norway and in Denmark. However, there were still differences 5 between the two datasets, especially for the spatial patterns of cropland before 1950. After 1950, the differences of spatial patterns of cropland between two datasets were much smaller than before. For Sweden, HYDE3.2 had more cropland area around lakes Vä nern, Vä ttern, Hjä lmaren and Mä laren in the ten time points. For Norway, from 1700 to 1810, HYDE3.2 showed that cropland fractions around Olsofjorden in Norway were more than 20%, but this 10 study showed there was few grid cells with cropland fraction more than 20% in Norway.
After 1880, cropland areas in the counties of Sogn & Fjordane, Hordaland and Rogaland in Southwestern Norway from HYDE3.2 were greater than those from this study. For Denmark, the negative differences in eastern Denmark between HYDE3.2 and this study indicated that HYDE3.2 underestimated the cropland area for the period 1700~1880. From 19101700~1880. From to 1950 HYDE3.2 revealed that the cropland area increased significantly in the whole Denmark and decreased gradually after 1950, which was contradictory to the spatial patterns of cropland distribution in this study. This study suggested that the area with high cropland fraction began represents land cultivated with long-term crops. In this study, based on the data sources, the definition of cropland is slightly different in different countries at different time periods. Thus, we tried to unify the definition of cropland in Scandinavia. In Sweden, in the periods of 10 1690-1810 and 1930-1950, 1875-1910, and 1980-2015, though "Åker", "Åker och annan odlad jord" and "Åk ermark" were used to record cropland, respectively, their meanings were similar to the cropland defined by FAO. In Denmark, the cropland was called "Ager" in 1690 and "Agerjord" in 1800, 1881 and 1998. Both "Ager" and "Agerjord" represent the total of areas which are annually sown with various crops and the resting land, and have the same 15 meaning as cropland defined by FAO. In the rest points of time, various types of agricultural land areas were recorded and we used the types which were included in cropland. In Norway, from 1690 to 1875, only volumes of seeds regarding different crops were recorded, which represented the land area for cultivation in the corresponding years. From 1910 to 1980, NSD classified the agricultural land according to the purpose of land use and recorded the 20 agricultural land area of each classification. After 1980, the data source only recorded the agricultural land area directly. As a result, in this study, cropland areas from 1690 to 1875 were lower and cropland areas after 1980 were higher than the real cropland areas in Norway.

Uncertainties in cropland area allocation
Uncertainties also exist in process of cropland area allocation. We used CCI-LC maps in 2000 25 and 2015 and combined 4 classes of rainfed cropland, irrigated or post-flooding cropland, mosaic cropland(>50%)/ natural vegetation (tree, shrub, herbaceous cover) (<50%) and mosaic natural vegetation (tree, shrub, herbaceous cover) (>50%) / cropland (<50%) to https://doi.org /10.5194/essd-2020-187 Open Access Earth System Science Data Discussions Preprint. Discussion started: 6 August 2020 c Author(s) 2020. CC BY 4.0 License. determine the maximum extent of cropland cover. However, since the maximum cropland cover extent map is Boolean, grid cells of mosaic cropland (>50%) and mosaic cropland (<50%) were considered as full of cropland area. More cropland area may be allocated in those grid cells. In Sweden and Denmark, the cropland area reached its maximum in 1950.
The spatial distribution of cropland around 1950 may have been greater than the maximum 5 cropland extent we selected, so that no cropland was allocated in the grids that should have had cropland. The difference between the total cropland area after cropland allocation and the total cropland area obtained from statistics is -7%~7% at the national scale, except for Norway in 1999. After allocation, the total cropland area of Norway in 1999 was even smaller than that in 2015, which shows the difference between remote sensing data and statistics. For All cropland data cover 1690,1750,1810,1875,1910,1930,1950,1980,1999

Conclusions
Based on the collected statistics of cropland area of each administrative unit, and using a 10 range of methods for data preprocessing and cropland area allocation, we developed the The comparison of our dataset and HYDE3.2 shows that in the whole of Scandinavia and in Sweden, the total cropland areas of HYDE3.2 are close to those of this study from 1700 to 2015. However, large differences are found in Norway and Denmark, especially before 1950.
The spatial patterns of cropland distribution in Scandinavia over the past 300 years from HYDE3.2 and this study are quite close and differences in cropland distribution over the past 10 300 years between HYDE3.2 and this study are mainly observed in regions which are highly cultivated.
Although the various definition of cropland in our data sources and errors in cropland area allocation brought uncertainties with our reconstruction, this study improves descriptions of historical cropland change in Scandinavia and the cropland dataset is an important reference 15 for better understanding of the complex climate system.