In-situ start and end of growing season dates of major European crop types from France and Bulgaria at a field level

Crop phenology data offer crucial information for crop yield estimation, agricultural management, and assessment of agroecosystems. Such information becomes more important in the context of increasing year-to-year climatic variability. The dataset provides in-situ crop phenology data (first leaves emergence and harvest date) of major European crops (wheat, corn, sunflower, rapeseed) from seventeen field study sites in Bulgaria and two in France. Additional information such as the sowing date, area of each site, coordinates, method and equipment used for phenophase data estimation, and photos of the France sites are also provided. The georeferenced ground-truth dataset provides a solid base for a better understanding of crop growth and can be used to validate the retrieval of phenological stages from remote sensing data.

Dataset link: In-situ crop phenology dataset from sites in Bulgaria and France (Original data)

Keywords:
Winter crop Summer crop Cover crop BBCH Phenology crops (wheat, corn, sunflower, rapeseed) from seventeen field study sites in Bulgaria and two in France.Additional information such as the sowing date, area of each site, coordinates, method and equipment used for phenophase data estimation, and photos of the France sites are also provided.The georeferenced ground-truth dataset provides a solid base for a better understanding of crop growth and can be used to validate the retrieval of phenological stages from remote sensing data.
© 2023 The Author(s).Published by Elsevier Inc.This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ) Specifications Table Subject Agricultural Sciences: Agronomy and Crop Science Specific subject area In-situ start and end of growing season data of numerous crop types at the field level Type of data Two Tables (text format), one with the crop phenology data and the additional information of each site and another with the description of each column variable of the crop phenology data.
Georeferenced plots from all sites (GeoJSON file format).
Open-format georeferenced plot data from all sites (CSV file format).
One file (text format), with support data for the georeferenced files.Photos from French sites (JPEG file format).How the data were acquired The sowing and harvest date were previously communicated to the researchers.The phenology in-situ data were acquired with systematic field observations by an agronomist, pictures taken by a NetCam SC IR (StarDot Technologies, Buena Park, CA) camera (phenocam) and digital photo camera (RICOH G700SE, RICOH International, Düsseldorf, Germany).The first leave emergence was determined in relation to the homogeneity or heterogeneity of the cover to which an average BBCH stage was attributed using a BBCH-scale which is traditionally used to identify the phenological development stages of plants.

Data format Raw data Description of data collection
We monitored seventeen production fields in Bulgaria (one growing season each) and two research fields in France (8 -10 growing seasons from 2015 -2022).The in-situ data were collected at a field level and consisted of sowing date, first leaves emergence (phenophase, BBCH10

Value of the Data
• In-situ start and end of growing season measurements at the field level are valuable because they offer direct ground-based observations on the timing and progression of crop growth stages [1] .• Improve the accuracy in identifying crop phenology stages [2] .
• Develop more accurate models and forecasts, assess agroecosystems, and understand how seasonal cycles may be altered by climate change [3] .• These data are highly valuable to ecologists, agronomists, and modelers, as the recent EU dataset [4] is deficient in providing such information that could be beneficial for their work.• When integrated with environmental and climatic data, crop phenology data can contribute to the study of climate impacts on agriculture, ultimately helping to optimize or adjust crop production and management [5] .• It can also serve as a reliable ground validation for remote sensing studies and provide valuable data to optimize phenology extraction models.

Objective
Our dataset [6] offers precise, field-level information on the timing of two critical growth stages: the start and end of the growing season.It is designed to advance research in land surface phenology (LSP) by enabling better understanding and prediction of crop phenology patterns.This knowledge is crucial for various purposes, including optimizing planting and harvesting schedules, managing crop health, and improving overall agricultural productivity.
The dataset serves as a resource for researchers, policymakers, and stakeholders in the agricultural sector.By analyzing the timing and duration of the growing season, we can enhance our understanding of crop phenology variations.This understanding allows for better predictions and informed decision-making regarding agricultural practices.
By maximizing the utility of existing data, we can contribute to the advancement of crop phenology research and support sustainable and resilient agricultural practices.

Data Description
We provide an in-situ crop phenology stages dataset from two different countries, Bulgaria and France, at a field level, see Fig. 1 .
In Bulgaria, we monitored seventeen fields, Fig. 2 , and in France two fields, Fig. 3 .The collected data set runs from the growing season 2015/2016 until 2018/2019 for Bulgarian field sites and 2015/2016 until 2021/2022 for France field sites, comprising a total of 35 observations (rows).This dataset contains information on phenology stages (sowing date, first leaves emergence/phenophase data, BBCH10-BBCH13, date, and harvest date), method and equipment used to access the phenophase, area, coordinates, crop type and the name of the    On the Zenodo repository, there is one metadata file (2_senseco_metadata_insitu_crop_pheno logy_Bulgaria_France.txt) with columns description of the actual dataset (1_senseco_data_ insitu_crop_phenology_Bulgaria_France.txt).
The main phenology information consists of sowing date, phenophase (BBCH10-13) data, date, and harvest date at a field level.Phenophase date corresponds to the first leave emergence characterized by BBCH10-13 distinguishable phenology stage.
The French sites are located in Southwestern France near Toulouse ( Fig. 3 ).The straight line distance between both sites is 12 kilometers.Both experimental plots are part of the Regional Spatial Observatory South West (OSR SW), the regional Zone Atelier Pyrénées-Garonne (ZA PY-GAR, [7] ), the national research infrastructure Critical Zone Observatories: Research and Applications (OZCAR; [8] ) and the Integrated Carbon Observation System (ICOS; [9] ) European network.
The FR-Lam field is part of an experimental dairy farm (Domaine de Lamothe, Ecole d'ingénieurs de Purpan) located in a plain where a winter wheat -irrigated maize crop rotation is performed.The crops are grown for feed and are fertilized with dairy excreta or slurry.The silage maize is harvested when the plant is still green, and the kernels are not yet ripe (i.e., between the middle and end of August).The FR-Lam site is located in a vulnerable zone, which means that it is subject to nitrate leaching.Within the framework of the 4th nitrate directive, issued from a European regulatory framework, farmers are obliged to cover the soil of agricultural plots located in these zones during the winter with the introduction of cover crops (i.e.white mustard, vetch, faba bean, and phacelia) between a winter crop and a summer crop.
The FR-Aur crop site is part of a grain farm located in a hilly area where a winter wheat -rapeseed -barley-sunflower crop rotation is performed.The soils of agricultural plots in this rugged landscape are often subject to erosion during periods of bare soil and intense rainfall.Therefore, to limit the loss of soil from their plots, farmers introduce intermediate cover crops during the winter between a winter and a summer crop.
On both French sites cover crops are crushed green and incorporated into the soil, not exported.
The climate on both sites is a temperate climate with oceanic and Mediterranean influences, with mild winters, rainy springs, and very hot summers with very low rainfall, followed by very sunny autumns.From a nearby MeteoFrance measurement station (Lherm-Muret), the mean annual rainfall calculated over the past 24 years was 617 ± 101 mm and the mean annual temperature was 13.7 ± 0.6 °C on both sites FR-Lam and FR-Aur [ 10 , 11 ].

Data acquisition
The parcel shape was acquired by manual outlining of each field in a Geographical Information System (GIS) software.The parcel area was computed by geometry calculator.The latitude and longitude of each field were acquired by generating a field centroid and extracting its coordinates.
Usually, with parcel-level phenological estimates, it is important to acknowledge that the parcel is rarely in a uniform phenological stage, as different parts of the parcel may exhibit varying growth rates.This variability requires the use of the 50% threshold rule to determine the overall phenological stage of the field.This widely employed approach in agricultural and ecological research provides a practical and standardized method for assessing the phenological stage of a field [13] .
In this study, the dates of emergence for all fields were determined using the 50% threshold rule.This rule is a simplification that allows for considering the entire parcel as being in a uniform phenological stage.While acknowledging the inherent variability within the parcel, applying the 50% threshold rule by both an agronomist in the field and through digital camera picture (phenocam and digital photo camera) analysis ensures a common collection protocol.
The agronomist's observation footprint extends from the immediate proximity of the edge of the plot to a distant view of the entire plot, encompassing both close-up and panoramic perspectives.The phenocam observation footprint covers approximately 5 to 10% of the plot.The image analysis entailed examining both the immediate close range and the distant areas of the plot.
The agronomist assessed the phenophase (emergence) during the moment of observation, while the analysis of photos involved evaluating the phenophase at the time of analysis.Despite these differences, the agronomist and images observations share sufficient similarities to enable comparability.The phenological development stage was identified using the BBCH-scale.
The harvest date was recorded as the start date of the harvest.In Bulgaria, crop data (sowing date, phenophase data, and harvest date) from seventeen fields were collected by visual observations, written in a paper notebook, and subsequently passed to an Excel file.The available data for each field is from one growing season and thirteen fields miss phenophase data or harvest date.
Bulgarian fields are production fields cultivated by companies or cooperatives.Each field has a dedicated agronomist that provides the sowing date, emergence phenophase (date, and harvest date.The agronomist frequently visits the field and evaluted the phenological stage of the field.In certain years, winter crops faced unfavorable growing conditions, such as drought during crop emergence.A decision is then made after the winter period whether to keep or destroy the crop.Consequently, for some fields that have been planted with winter crops, only a portion of the crop may survive, with the rest being destroyed.The dataset provides information on the shape and area of the remaining crop that has been kept until harvest.The collection of the dataset took place during Dessislava Ganeva PhD research, which had a different objective than phenology monitoring.As a result, some fields within the dataset lack emergence dates or harvest dates.Nevertheless, we made the decision to include these plots with incomplete data in the proposed dataset due to the valuable information they still provide. In France, crop data (sowing date and harvest date) from 2 fields (FR-Aur and FR-Lam) were collected through visual observation, identical to the Bulgaria sites.The phenophase data was collected by field observations, phenocam picture and/or digital photo camera picture analysis.The exception is for the FR-Lam site during the season 2017/2018 and 2019/2020 where the phenophase data was collected only through field observations, identical to the Bulgaria sites.The available data for both fields cover numerous growing seasons.

Fig. 1 .
Fig. 1.The location of study fields (yellow dots) where the in-situ crop phenology stages were determined and acquired.Background -google satellite image.

Fig. 2 .
Fig.2.Shape of the selected seventeen field study sites in Bulgaria, P0, P01, P02, P1, P2, P3, P4, P13, P14, P15, P17, P18, P19, P20, P22, P23, and P24 with the corresponding elevation above sea level (a.s.l) in meters (m) and the area in hectares (ha).Each site is outlined.On the right are the map of Bulgaria and the name of the region where the sites are located.The scale bar in the top right refers to the image of the country and, in the bottom right it refers to the images of the field study sites.

Fig. 3 .
Fig. 3. Shape of the selected field study sites in France, Auradé (FR-Aur) and Lamasquère (FR-Lam) on the left with the corresponding elevation above sea level (a.s.l) in meters (m) and the area in hectares (ha).Each site is outlined.On the top right are the map of France and the name of the city where the sites are located.The scale bar on the right of the figure refers to the image of the country and, in the left bottom it refers to the images of the field study sites.

Table 1
Field identification and in-situ data structure: variables, type, descriptions, and unit or category at field level.Missing values represented by 'NA' (txt format).
associated photos of the France sites and the georeference data of each field site (file name: 1_senseco_data_insitu_crop_phenology_Bulgaria_France.txt).See the details of the file in Table1.The photos of the France field sites are in JPG format.All field study sites from both countries are georeferenced in GeoJSON and CSV format.There is a metadata file (metadata_geojson_csv.txt)with support information for the field study sites of both countries related to the GeoJSON and CSV format files.