GCI present an annual dynamic global cropping intensity dataset covering the period from 2001 to 2019 at a 250-m resolution with an average overall accuracy of 89%. We used the enhanced vegetation index (EVI) of MOD13Q1 as the database via a sixth-order polynomial function to calculate the cropping intensity. The global cropping intensity dataset was packaged in the GeoTIFF file type, with the quality control band in the same format. The dataset fills the vacancy of medium-resolution, global-scale annual cropping intensity data and provides an improved map for further global yield estimations and food security analyses. GCI and GCI_QC maps at a 250-m resolution were provided for the entire world from 2001 to 2019. The datasets and their validation samples are available at the figshare repository in GeoTIFF format and provided in the GCS_WGS_1984 spatial reference system. The global cropping intensity maps contain values of 0, 1, 2 and 3, representing none, single, double, and triple cropping, respectively. The QC band maps also contain values of 0, 1, 2 and 3, representing best, good, fair, and poor pixels, respectively. The dataset extends from 70° N to 60° S latitude and from 180° W to 180° E longitude, excluding Greenland and Antarctica. The maps can be visualized and analysed in ArcGIS, QGIS, or similar software.

Since late 2020, one of the worst wars has been raging in Tigray, Ethiopia's northernmost region. A humanitarian tragedy has been caused by the fighting (Dedefo Bedaso 2021; Annys et al. 2021). Intense fighting occurred throughout the whole region, and looting and damage were rampant. Farmers were harvesting their crops in the middle of a desert locust infestation when the conflict began in late 2020. To record war impacts, commonly, direct expenses or losses at a particular period are quantified (Lindgren 2004). Post-conflict damage assessments typically concentrate on losses to businesses, services, infrastructure, and facilities in cities, even though the primary source of income in developing countries is the agricultural sector. Even when agricultural evaluations are done, they mostly focus on crop losses and ignore how wars affect land management. In Tigray's small-scale family farms, which use a permanent farming system based on cereals, oxen are utilized for traction (Westphal 1975). Crop agriculture has been practiced in Tigray for at least three thousand years (D'Andrea 2008; Blond et al. 2018), allowing for the gradual improvement of the agricultural system, including considerable farmers' understanding of the procedures involved in seed selection and of land suitability (Fetien Abay et al. 2008). The indigenous knowledge (sensu Bruchac 2018) also includes a broad vocabulary for different soil types (Nyssen et al. 2019), and the capacity to interpret the rainy season when selecting the crop to be sown (Frankl et al. 2013). A significant degree of equality in the extent of landholdings has resulted from the strengthening of the egalitarian land tenure system during the 1980s (Hendrie 1999). In the study region, a typical household uses two or three farmland pieces totalling less than a hectare. The ownership and management of grasslands, rangelands, and woodlands are communal (Nyssen et al. 2008). In the first half of 2021, armed forces of the Ethiopian government and from Eritrea as well as from the neighboring Amhara region were engaged in warfare against the forces of Tigray's regional government; in the second half of the year, warfare was essentially outside of Tigray, more to the south, while Tigray itself was submitted to a blockade with all telecommunication and lifelines to the outside world cut (Pellet 2021; Gayim 2021; Ramos 2021), a blockade that continued into 2022. The near-absence of economic activities, combined with limited food stocks and restricted humanitarian access resulted in 70% of the population experiencing starvation (sensu Stratton et al. 2003), i.e. high levels of acute food insecurity and excess mortality (Plaut 2021; Istratii 2021; Teklehaymanot G Weldemichel 2021; Oxford Analytica 2021; Devi 2021; Müller and Read 2021). The famine was worst from September to December 2021, as it took up to December before the years' poor harvest could be consumed REF; the lean period (also called “lean season”, “hunger season”) has been very severe. The lean season is the time in between finishing the last food that people had at hand and starting to consume the new harvest (Hirvonen et al. 2016). Farmers' main goals in these dire circumstances were to attempt to generate a better harvest in 2022 and, despite everything, to try and survive another year. We offer field data obtained by the end of August 2022, which were evaluated to determine the percentage of Tigray's land that was seeded on schedule, the types of crops sown, and the condition of these crops. Despite difficult living and travel conditions, the agricultural status in some of Tigray's reachable districts was examined for the 2022 growing season. A team of geographers visited 262 agricultural plots in an area indicative of the region's diverse bio-physical circumstances, including elevation (plots ranged from 1931 to 2600 meters above sea level), lithology, soil type, rainfall patterns, and hence cropping strategies (Alemtsehay Tsegay et al. 2019; Nyssen et al. 2019). Other land uses, such as irrigated land, grassland, barren land, bushland, and forest, were left out of the analysis. We visited ecoregions with different biophysical and agro-ecological characteristics along main roads in six districts between 24 and 29 August 2022: Tsa'ida Imba, Kilte Awula'ilo (especially croplands on the outskirts of Wukro's urban district), Dogu'a Tembien (surroundings of Hagere Selam), Samre, Hintalo (particularly Addi Gudom), and Inderta (Aynalem and Didiba). The investigations typically took place in the wider surroundings of small towns, as transect walks, observing and talking to farmers present on the land. Participatory monitoring was used to collect data for each cropland, which included recording the crop type, a group assessment of the crop's status according to local standards (good, medium, bad, failed; taking into account growth features such as plant height, greenness and density, ear length, homogeneity in crop stand), observations of whether or not neighboring farmers cropped in block, and a semi-structured interview with the farmer or a group discussion, addressing among others the use of fertilizer (Van De Fliert et al. 2000; Nyumba et al. 2018; Young and Hinton 1996). Aside from the usual crop evaluation, emphasis was paid to block wise cropping with adjacent farmers since, like three-field systems, this practice is an indicative of an internally well-organized community, and hints to a superior yield forecast as it prevents disruptions (Nyssen et al. 2008; Hopcroft 1994; Ruthenberg 1980). Data have been collected in such a way that homogeneous areas of at least 30 m x 30 m are represented, so that they can serve as calibration and validation points in remote sensing analysis. According to descriptive statistics from the dataset, at the end of August, 15% of the monitored farm parcels had been left fallow, meaning no crops had been planted (40 plots out of 262). During a similar monitoring in 2021 (Tesfaalem Ghebreyohannes et al. 2022a), 21% of the monitored lands were fallowed. However, 7 percent of the fallow plots had no weeds, indicating that the ground had been ploughed but not seeded. A further 4% of the plots were planted with flax or niger seed, which is often used to improve fallow soil quality rather than crop output. Among the cultivated parcels, 104 plots (40%) were planted with wheat, barley, or a mixture of both (hanfets) (49% in 2021), while 84 plots (32%) were planted with tef (26% in 2021). Only 1% of the land was planted with maize, and another 1% with sorghum (6% and 4% in 2021). In the plots containing crops that were examined, 46% had been seeded in block, in collaboration with the owners of surrounding lands (40% in 2021). Wheat and barley (54%) as well as tef (52%) were seeded in blocks. Three quarter (76 percent) of the wheat and barley fields were in good or medium condition. Seventy-one percent of the tef lands were in poor condition (67% in 2021). Overall, the crop stands improved slightly over those of the very bad year 2021 (Tesfaalem Ghebreyohannes et al. 2022b), and there was less fallowing. Fertilizer was used on only 56 of the 222 sampled plots with crop: on these lands, at least some mineral fertilizer was administered at sowing, after crop emergence, or manure was applied. Due to a shortage of fertilizers, farmers frequently applied insufficient amounts. Mineral fertilizer was used exclusively for cereal production. A significant issue was the farmers' inappropriate use of potassium fertilizer, which led to crop burn, particularly in Tsa'ida Imba and Samre. Overall, and adopting a very low threshold, 34% of the analyzed lands were fallowed or are expected to provide a very poor crop harvest, while 66% of the sampled fields are promising and would yield medium or excellent crops.

'Sandy_soil_DB' - Database (see files tab) and a 'R shiny Sandbox app' (see link in services tab) that allows the user to explore yield responses to treatments for increasing production on sandy soils in low and medium rainfall areas of the Southern Region.

The database is a collection of trial results in the Southern cropping region from the Grains Research & Development Corporation CSP00203 project reporting on yield results, but also contains plant responses, such as plant establishment and biomass for a selection of sites / years. The aim of the Sandy Soils project is to boost crop yields by increasing crop water in underperforming sandy soils by improving the diagnosis and management of constraints. Crop water-use and yields on sandy soils are commonly limited by a range of soil constraints that reduce root growth. Constraints can include a compacted or hard-setting layer preventing root proliferation, a water-repellent surface layer causing poor crop establishment, soil pH issues (both acidity and alkalinity) and/or poor nutrient supply. The trails were established between 2014 and 2021 and included a range of deep ripping (30-60cm deep), spading, inclusion ripping and/or inversion ploughing approaches, with/without additional amendments (fertiliser, N-rich hay, chicken manure, clay), in total the project consisted of 136 site.years of amelioration trial data and 1 site.year of seeder-strategy trial data, across 24 sites in Southern grain region. Additional site metadata includes soil constraints and climate data. The database provides statistical analysis of the yield results, which was undertaken on a site-by-year basis and site-by-cumulative years.

The App is an interactive web application that allows the user to interrogate and visualise trial results alongside site climate and soil constraint information. This helps grain growers evaluate the outcomes of various amelioration options in the context of their own soils and climate. A series of fact sheets relating to soil constraint identification and machinery selection and optimisation support the user to view the most useful results and to identify the most suitable management options.

Lineage: The trial program consisted of 24 trials across the Southern Cropping Region of Australia in the low to medium rainfall environment. All sites had constraints of one or a combination of; high soil strength, water repellence or poor soil fertility. Research experiments were established between 2014 and 2020 and included a range of deep ripping, spading and/or ploughing approaches, with/without additional amendments (fertiliser, N-rich hay, chicken manure, clay). All trials monitored the impact of amelioration on crop growth and yield. More detailed description of the experimental approach is outlined in Unkovich et al. (2023).

The database provides soil constraints, plant responses and site climate and metadata for all trial sites. The database and App include data analysis outputs at the level of the site and are consolidated by treatment. Statistical analysis is reported for site by year and site by cumulative yields (for all the years the trial was run) using one-way ANOVA analysis and multiple comparisons of treatments by means of LSD and a grouping of treatments. The level by alpha is set at 0.1. The analysis was performed in R studio using the stats and agricolae packages.

StreetSurfaceVis

StreetSurfaceVis is an image dataset containing 9,122 street-level images from Germany with labels on road surface type and quality. The CSV file streetSurfaceVis_v1_0.csv contains all image metadata and four folders contain the image files. All images are available in four different sizes, based on the image width, in 256px, 1024px, 2048px and the original size.Folders containing the images are named according to the respective image size. Image files are named based on the mapillary_image_id.

You can find the corresponding publication here: StreetSurfaceVis: a dataset of crowdsourced street-level imagery with semi-automated annotations of road surface type and quality

Image metadata

Each CSV record contains information about one street-level image with the following attributes:

mapillary_image_id: ID provided by Mapillary (see information below on Mapillary)

user_id: Mapillary user ID of contributor

user_name: Mapillary user name of contributor

captured_at: timestamp, capture time of image

longitude, latitude: location the image was taken at

train: Suggestion to split train and test data. True for train data and False for test data. Test data contains data from 5 cities which are excluded in the training data.

surface_type: Surface type of the road in the focal area (the center of the lower image half) of the image. Possible values: asphalt, concrete, paving_stones, sett, unpaved

surface_quality: Surface quality of the road in the focal area of the image. Possible values: (1) excellent, (2) good, (3) intermediate, (4) bad, (5) very bad (see the attached Labeling Guide document for details)

Image source

Images are obtained from Mapillary, a crowd-sourcing plattform for street-level imagery. More metadata about each image can be obtained via the Mapillary API . User-generated images are shared by Mapillary under the CC-BY-SA License.

For each image, the dataset contains the mapillary_image_id and user_name. You can access user information on the Mapillary website by https://www.mapillary.com/app/user/ and image information by https://www.mapillary.com/app/?focus=photo&pKey=

If you use the provided images, please adhere to the terms of use of Mapillary.

Instances per class

Total number of images: 9,122

excellent good intermediate bad very bad

asphalt 971 1697 821

246

concrete 314 350 250

58

paving stones 385 1063 519

70

sett

129 694

540

unpaved

-

326 387 303

For modeling, we recommend using a train-test split where the test data includes geospatially distinct areas, thereby ensuring the model's ability to generalize to unseen regions is tested. We propose five cities varying in population size and from different regions in Germany for testing - images are tagged accordingly.

Number of test images (train-test split): 776

Inter-rater-reliablility

Three annotators labeled the dataset, such that each image was annotated by one person. Annotators were encouraged to consult each other for a second opinion when uncertain.1,800 images were annotated by all three annotators, resulting in a Krippendorff's alpha of 0.96 for surface type and 0.74 for surface quality.

Recommended image preprocessing

As the focal road located in the bottom center of the street-level image is labeled, it is recommended to crop images to their lower and middle half prior using for classification tasks.

This is an exemplary code for recommended image preprocessing in Python:

from PIL import Imageimg = Image.open(image_path)width, height = img.sizeimg_cropped = img.crop((0.25 * width, 0.5 * height, 0.75 * width, height))

License

CC-BY-SA

Citation

If you use this dataset, please cite as:

Kapp, A., Hoffmann, E., Weigmann, E. et al. StreetSurfaceVis: a dataset of crowdsourced street-level imagery annotated by road surface type and quality. Sci Data 12, 92 (2025). https://doi.org/10.1038/s41597-024-04295-9

@article{kapp_streetsurfacevis_2025, title = {{StreetSurfaceVis}: a dataset of crowdsourced street-level imagery annotated by road surface type and quality}, volume = {12}, issn = {2052-4463}, url = {https://doi.org/10.1038/s41597-024-04295-9}, doi = {10.1038/s41597-024-04295-9}, pages = {92}, number = {1}, journaltitle = {Scientific Data}, shortjournal = {Scientific Data}, author = {Kapp, Alexandra and Hoffmann, Edith and Weigmann, Esther and Mihaljević, Helena}, date = {2025-01-16},}

This is part of the SurfaceAI project at the University of Applied Sciences, HTW Berlin.

• Prof. Dr. Helena Mihajlević- Alexandra Kapp- Edith Hoffmann- Esther Weigmann

Contact: surface-ai@htw-berlin.de

https://surfaceai.github.io/surfaceai/

Funding: SurfaceAI is a mFund project funded by the Federal Ministry for Digital and Transportation Germany.