Geology Reports⌕ Search

SEARCH · Geology Reports

Results for “Precision Agriculture”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 91 records · Page 5Linked to original sources
USGS2021

Miniature temperature data loggers increase precision and reduce bias when estimating the daily survival rate for bird nests

Demographic studies of many bird species are challenging because their nests are cryptic, resulting in few nests being found. To maximize statistical power, methods are needed that minimize disturbance while yielding as much information per nest as possible. One way to meet these objectives is to use miniature thermal data loggers to precisely date nest fates. Our objectives, therefore, were to (1) examine the possible effect of thermal data loggers on nest success through hatching by grass- and shrub-nesting songbirds that differed in their parasite egg-accepting and -rejecting behavior, (2) examine the effect of using daily temperature data versus less frequent nest-visit data on statistical power, bias, and precision when estimating the daily survival rate (DSR) for nests, and (3) compare these two approaches using a simulation study and field data. We monitored the survival of nests located in agricultural landscapes and used a binomial logistic regression with main effects for data-loggers and parasite-accepting or -rejecting status and their interaction. We also compared maximum likelihood–derived DSR for differences in estimated rates, precision, and sample sizes with both data collected in the field and simulated with varying sample sizes and visit frequencies. We found no evidence that thermal data loggers had any effect on hatching rates either for all species or for parasite egg-accepting and -rejecting species, separately. Both our simulation and analysis of real nest data indicated that use of data loggers increased the statistical power from each nest studied by increasing effective sample sizes and precision of DSR estimates compared to in-person visits. We also found a negative bias in DSR estimates with longer visit intervals, which use of data-loggers removed. Both the results of simulated- and field-data analyses suggest that future studies of nest survival can be improved by automated nest monitoring by removing a source of bias and providing more time to find additional nests.

Iowa↗
USGS2011

Evaluation of an index of biotic integrity approach used to assess biological condition in western U.S. streams and rivers at varying spatial scales

Consistent assessments of biological condition are needed across multiple ecoregions to provide a greater understanding of the spatial extent of environmental degradation. However, consistent assessments at large geographic scales are often hampered by lack of uniformity in data collection, analyses, and interpretation. The index of biotic integrity (IBI) has been widely used in eastern and central North America, where fish assemblages are complex and largely composed of native species, but IBI development has been hindered in the western United States because of relatively low fish species richness and greater relative abundance of alien fishes. Approaches to developing IBIs rarely provide a consistent means of assessing biological condition across multiple ecoregions. We conducted an evaluation of IBIs recently proposed for three ecoregions of the western United States using an independent data set covering a large geographic scale. We standardized the regional IBIs and developed biological condition criteria, assessed the responsiveness of IBIs to basin-level land uses, and assessed their precision and concordance with basin-scale IBIs. Standardized IBI scores from 318 sites in the western United States comprising mountain, plains, and xeric ecoregions were significantly related to combined urban and agricultural land uses. Standard deviations and coefficients of variation revealed relatively low variation in IBI scores based on multiple sampling reaches at sites. A relatively high degree of corroboration with independent, locally developed IBIs indicates that the regional IBIs are robust across large geographic scales, providing precise and accurate assessments of biological condition for western U.S. streams. ?? Copyright by the American Fisheries Society 2008.

Transactions of the American Fisheries Society↗
USGS2026

Improving building footprint extraction using NAIP and 3DEP lidar derived features with deep learning

Accurate building footprint extraction is critical for applications ranging from population estimation to disaster management. Although optical imagery provides detailed spectral information, it often struggles with shadows, occlusions, and background clutter in dense urban environments. Lidar data, by contrast, offer precise elevation and structural attributes but face challenges such as variable point density and noise. This study integrates multispectral imagery from the U.S. Department of Agriculture (USDA) National Agriculture Imagery Program (NAIP) with lidar-derived feature height and intensity from the U.S. Geological Survey (USGS) 3D Elevation Program (3DEP) to improve footprint extraction using a U-Net–based deep learning model. A six-band input stack (RGB, near-infrared, height, intensity) was developed, normalized, and tiled for training and evaluation against Microsoft Global Building Footprints (GBF). Results from the Houston, TX test site show that the six-band model achieved a precision of 0.86, recall of 0.88, F1 score of 0.87, and Intersection-over-Union (IoU) of 0.76, consistently outperforming four-band baselines by reducing false positives while maintaining sensitivity. Predictions on withheld Houston tiles confirmed strong within-region generalization, yielded a precision of 0.78, recall of 0.81, F1 score of 0.79, and IoU of 0.66. Qualitative analysis further revealed limitations stemming from both training label quality and vegetation–building confusion. These findings demonstrate the complementary value of integrating spectral and structural information for robust building footprint extraction and how domain adaptation strategies can be used to enhance cross-regional transferability.

Texas↗
USGS2012

Hydrological information products for the Off-Project Water Program of the Klamath Basin Restoration Agreement

The Klamath Basin Restoration Agreement (KBRA) was developed by a diverse group of stakeholders, Federal and State resource management agencies, Tribal representatives, and interest groups to provide a comprehensive solution to ecological and water-supply issues in the Klamath Basin. The Off-Project Water Program (OPWP), one component of the KBRA, has as one of its purposes to permanently provide an additional 30,000 acre-feet of water per year on an average annual basis to Upper Klamath Lake through "voluntary retirement of water rights or water uses or other means as agreed to by the Klamath Tribes, to improve fisheries habitat and also provide for stability of irrigation water deliveries." The geographic area where the water rights could be retired encompasses approximately 1,900 square miles. The OPWP area is defined as including the Sprague River drainage, the Sycan River drainage downstream of Sycan Marsh, the Wood River drainage, and the Williamson River drainage from Kirk Reef at the southern end of Klamath Marsh downstream to the confluence with the Sprague River. Extensive, broad, flat, poorly drained uplands, valleys, and wetlands characterize much of the study area. Irrigation is almost entirely used for pasture. To assist parties involved with decisionmaking and implementation of the OPWP, the U.S. Geological Survey (USGS), in cooperation with the Klamath Tribes and other stakeholders, created five hydrological information products. These products include GIS digital maps and datasets containing spatial information on evapotranspiration, subirrigation indicators, water rights, subbasin streamflow statistics, and return-flow indicators. The evapotranspiration (ET) datasets were created under contract for this study by Evapotranspiration, Plus, LLC, of Twin Falls, Idaho. A high-resolution remote sensing technique known as Mapping Evapotranspiration at High Resolution and Internalized Calibration (METRIC) was used to create estimates of the spatial distribution of ET. The METRIC technique uses thermal infrared Landsat imagery to quantify actual evapotranspiration at a 30-meter resolution that can be related to individual irrigated fields. Because evaporation uses heat energy, ground surfaces with large ET rates are left cooler as a result of ET than ground surfaces that have less ET. As a consequence, irrigated fields appear in the Landsat images as cooler than nonirrigated fields. Products produced from this study include total seasonal and total monthly (April-October) actual evapotranspiration maps for 2004 (a dry year) and 2006 (a wet year). Maps showing indicators of natural subirrigation were also provided by this study. "Subirrigation" as used here is the evapotranspiration of shallow groundwater by plants with roots that penetrate to or near the water table. Subirrigation often occurs at locations where the water table is at or above the plant rooting depth. Natural consumptive use by plants diminishes the benefit of retiring water rights in subirrigated areas. Some agricultural production may be possible, however, on subirrigated lands for which water rights are retired. Because of the difficulty in precisely mapping and quantifying subirrigation, this study presents several sources of spatially mapped data that can be used as indicators of higher subirrigation probability. These include the floodplain boundaries defined by stream geomorphology, water-table depth defined in Natural Resources Conservation Service (NRCS) soil surveys, and soil rooting depth defined in NRCS soil surveys. The two water-rights mapping products created in the study were "points of diversion" (POD) and "place of use" (POU) for surface-water irrigation rights. To create these maps, all surface-water rights data, decrees, certificates, permits, and unadjudicated claims within the entire 1,900 square mile study area were aggregated into a common GIS geodatabase. Surface-water irrigation rights within a 5-mile buffer of the study area were then selected and identified. The POU area was then totaled by water right for primary and supplemental water rights. The maximum annual volume (acre-feet) allowed under each water right also was calculated using the POU area and duty (allowable annual irrigation application in feet). In cases where a water right has more than one designated POD, the total volume for the water right was equally distributed to each POD listed for the water right. Because of this, mapped distribution of diversion rates for some rights may differ from actual practice. Water-right information in the map products was from digital datasets obtained from the Oregon Water Resources Department and was, at the time acquired, the best available compilation of water-right information available. Because the completeness and accuracy of the water-right data could not be verified, users are encouraged to check directly with the Oregon Water Resources Department where specific information on individual rights or locations is essential. A dataset containing streamflow statistics for 72 subbasins in the study area was created for the study area. The statistics include annual flow durations (5-, 10-, 25-, 50-, and 95-percent exceedances) and 7-day, 10-year (7Q10) and 7-day, 2-year (7Q2) low flows, and were computed using regional regression equations based on measured streamflow records in the region. Daily streamflow records used were adjusted as needed for crop consumptive use; therefore the statistics represent streamflow under more natural conditions as though irrigation diversions did not exist. Statistics are provided for flow rates resulting from streamflow originating from within the entire drainage area upstream of the subbasin pour point (referring to the outlet of the contributing drainage basin). The statistics were computed for the purpose of providing decision makers with the ability to estimate streamflow that would be expected after water conservation techniques have been implemented or a water right has been retired. A final product from the study are datasets of indicators of the potential for subsurface return flow of irrigation water from agricultural areas to nearby streams. The datasets contain information on factors such as proximity to surface-water features, geomorphic floodplain characteristics, and depth to water. The digital data, metadata, and example illustrations for the datasets described in this report are available on-line from the USGS Water Resources National Spatial Data Infrastructure (NSDI) Node Website http://water.usgs.gov/lookup/getgislist or from the U.S. Government website DATA.gov at http://www.data.gov with links provided in a Microsoft® Excel® workbook in appendix A.

California;Oregon↗
USGS2004

Using magnetic susceptibility to facilitate more rapid, reproducible and precise delineation of hydric soils in the midwestern USA

Standard field indicators, currently used for hydric soil delineations [USDA-NRCS, 1998. Field indicators of hydric soils in the United States, Version 4.0. In: G.W. Hurt et al. (Ed.), United States Department of Agriculture-NRCS, Fort Worth, TX], are useful, but in some cases, they can be subjective, difficult to recognize, or time consuming to assess. Magnetic susceptibility (MS) measurements, acquired rapidly in the field with a portable meter, have great potential to help soil scientists delineate and map areas of hydric soils more precisely and objectively. At five sites in Illinois (from 5 to 15 ha in area) with contrasting soil types and glacial histories, the MS values of surface soils were measured along transects, and afterwards mapped and contoured. The MS values were found to be consistently higher in well-drained soils and lower in hydric soils, reflecting anaerobic deterioration of both detrital magnetite and soil-formed ferrimagnetics. At each site, volumetric MS values were statistically compared to field indicators to determine a critical MS value for hydric soil delineation. Such critical values range between 22??10-5 and 33??10-5 SI in silty loessal or alluvial soils in Illinois, but are as high as 61??10-5 SI at a site with fine sandy soil. A higher magnetite content and slower dissolution rate in sandy soils may explain the difference. Among sites with silty parent material, the lowest critical value (22??10-5 SI) occurs in soil with low pH (4.5-5.5) since acidic conditions are less favorable to ferrimagnetic mineral neoformation and enhance magnetite dissolution. Because of their sensitivity to parent material properties and soil pH, critical MS values must be determined on a site specific basis. The MS of studied soil samples (0-5 cm depth) is mainly controlled by neoformed ultrafine ferrimagnetics and detrital magnetite concentrations, with a minor contribution from anthropogenic fly ash. Neoformed ferrimagnetics are present in all samples but, based on high ??FD% (???5% to 10%), are most prevalent in high pH Mollisols of northeastern Illinois. Scanning electron microscope images display significantly more detrital magnetite alteration in hydric soils, substantiating that reductive dissolution of magnetite (aided by microorganisms) is a primary cause for lower MS. Fly ash comprises 8-50% of the >5 ??m strongly magnetic particles and typically accounts for 5-15% of the total MS signal. The proportion of fly ash in >5 ??m strongly magnetic fractions is greater in hydric soils because of lower natural magnetite contents, possibly combined with historical topsoil accumulation in lower landscapes. Magnetic fly ash particles are also more altered in low MS soils, implying that significant magnetite dissolution can occur in less than 150 years. ?? 2004 Elsevier B.V. All rights reserved.

Catena↗
USGS2019

Agricultural cropland extent and areas of South Asia derived using Landsat satellite 30-m time-series big-data using random forest machine learning algorithms on the Google Earth Engine cloud

The South Asia (India, Pakistan, Bangladesh, Nepal, Sri Lanka and Bhutan) has a staggering 900 million people (~43% of the population) who face food insecurity or severe food insecurity as per United Nations, Food and Agriculture Organization’s (FAO) the Food Insecurity Experience Scale (FIES). The existing coarse-resolution (>250-m) cropland maps lack precision in geo-location of individual farms and have low map accuracies. This also results in uncertainties in cropland areas calculated from such products. Thereby, the overarching goal of this study was to develop high spatial resolution (30-m or better) baseline cropland extent product of South Asia for the year 2015 using Landsat satellite time-series big-data and machine learning algorithms (MLAs) on the Google Earth Engine (GEE) cloud computing platform. To eliminate the impact of clouds, ten time-composited Landsat bands (blue, green, red, NIR, SWIR1, SWIR2, Thermal, EVI, NDVI, NDWI) were derived for each of the 3 time-periods over 12 months (monsoon: Julian days 151-300; winter: Julian days 301-365 plus 1-60; and summer: Julian days 61-150), taking the every 8-day data from Landsat-8 and 7 for the years 2013-2015, for a total of 30-bands plus global digital elevation model (GDEM) derived slope band. This 31-band mega-file big data-cube was composed for each of the 5 agro-ecological zones (AEZ’s) of South Asia and formed a baseline data for image classification and analysis. Knowledge-base for the Random Forest (RF) MLAs were developed using spatially well spread-out reference training data (N=2179) in 5 AEZs. Classification was performed on GEE for each of the 5 AEZs using well-established knowledge-based and RF MLAs on the cloud. Map accuracies were measured using independent validation data (N=1185). The survey showed that the South Asia cropland product had a producer’s accuracy of 89.9% (errors of omissions of 10.1%), user’s accuracy of 95.3% (errors of commission of 4.7%) and an overall accuracy of 88.7%. The National and sub-national (districts) areas computed from this cropland extent product explained 80-96% variability when compared with the National statistics of the South Asian Countries. The full resolution imagery can be viewed at full-resolution, by zooming-in to any location in South Asia or the world, at www.croplands.org and the cropland products of South Asia downloaded from The Land Processes Distributed Active Archive Center (LP DAAC) of National Aeronautics and Space Administration (NASA) and the United States Geological Survey (USGS): https://lpdaac.usgs.gov/products/gfsad30saafgircev001/

GIScience and Remote Sensing↗
USGS2013

Biochar effect on maize yield and soil characteristics in five conservation farming sites in Zambia

Biochar addition to agricultural soils can improve soil fertility, with the added bonus of climate change mitigation through carbon sequestration. Conservation farming (CF) is precision farming, often combining minimum tillage, crop rotation and residue retention. In the present farmer-led field trials carried out in Zambia, the use of a low dosage biochar combined with CF minimum tillage was tested as a way to increase crop yields. Using CF minimum tillage allows the biochar to be applied to the area where most of the plant roots are present and mirrors the fertilizer application in CF practices. The CF practice used comprised manually hoe-dug planting 10-L sized basins, where 10%–12% of the land was tilled. Pilot trials were performed with maize cob biochar and wood biochar on five soils with variable physical/chemical characteristics. At a dosage as low as 4 tons/ha, both biochars had a strong positive effect on maize yields in the coarse white aeolian sand of Kaoma, West-Zambia, with yields of 444% ± 114% ( p = 0.06) and 352% ± 139% ( p = 0.1) of the fertilized reference plots for maize and wood biochar, respectively. Thus for sandy acidic soils, CF and biochar amendment can be a promising combination for increasing harvest yield. Moderate but non-significant effects on yields were observed for maize and wood biochar in a red sandy clay loam ultisol east of Lusaka, central Zambia (University of Zambia, UNZA, site) with growth of 142% ± 42% ( p > 0.2) and 131% ± 62% ( p > 0.2) of fertilized reference plots, respectively. For three other soils (acidic and neutral clay loams and silty clay with variable cation exchange capacity, CEC), no significant effects on maize yields were observed ( p > 0.2). In laboratory trials, 5% of the two biochars were added to the soil samples in order to study the effect of the biochar on physical and chemical soil characteristics. The large increase in crop yield in Kaoma soil was tentatively explained by a combination of an increased base saturation (from <50% to 60%–100%) and cation exchange capacity (CEC; from 2–3 to 5–9 cmol/kg) and increased plant-available water (from 17% to 21%) as well as water vapor uptake (70 mg/g on maize cob biochar at 50% relative humidity).

Agronomy Journal↗
USGS2020

Mapping croplands of Europe, Middle East, Russia, and Central Asia using Landsat 30-m data, machine learning algorithms and Google Earth Engine

Accurate and timely information on croplands is important for environmental, food security, and policy studies. Spatially explicit cropland datasets are also required to derive information on crop type, crop yield, cropping intensity, as well as irrigated areas. Large area defined as continental to global cropland mapping is challenging due to differential manifestation of croplands, wide range of cultivation practices and limited reference data availability. This study presents the results of a cropland extent mapping of 64 Countriescovering large parts of Europe, Middle East, Russia and Central Asia. To cover such a vast area, roughly 160,000 Landsat scenes from 3,351 footprints between 2014 and 2016 were processed within the Google Earth Engine (GEE) cloud-platform. We used the pixel-based supervised Random Forest (RF) machine learning algorithm with a set of satellite data inputs capturing diverse spectral, temporal and topographical characteristics across twelve agroecological zones (AEZs). The reference data to train the classification model were collected from very high spatial resolution imagery (VHRI) and ancillary datasets. The result is a binary map showing cultivated/non-cultivated areas ca. 2015. The map produced an overall accuracy of 94 percent with roughly 14 percent omission and commission errors for the cropland class based on a large set of independent validation samples. The map suggests the entire study area has a total 546 million hectares (Mha) of croplands occupying 18 percent of the land area. Comparison between national cropland area estimates from United Nations Food and Agricultural Organizations (FAO) and those derived from this work also showed an R-square value of 0.95. For the entire Landsat-derived 30-m product the overall accuracy was 93.8% with cropland class providing producers accuracy of 86.5% (errors of omissions = 13.5%) and users accuracy of 85.7% (errors of commissions = 14.3%). This Landsat-derived 30-m cropland product (GFSAD30) provided 10-30% greater cropland areas compared to UN FAO in the 64 Countries. Finally, the map-to-map comparison between GFSAD30 with several other cropland products revealed that the best similarity matrix was with the 30m global land cover (GLC30) product providing an overall accuracy of 88.8 percent (Kappa 0.7) with producers cropland similarity of 89.2 percent (errors of omissions = 10.8%) and users cropland similarity of 81.8 percent (errors of commissions = 8.1%). GFSAD30 captured the missing croplands in GLC30 product around significantly irrigated agricultural areas in Germany and Belgium and rainfed agriculture in Italy. This study also established that the real strength of GFSAD30 product, compared to other products, were in: 1. Identifying precise location of croplands, and 2. Capturing fragmented croplands. The cropland extent map dataset is available through NASAs Land Processes Distributed Active Archive Center (LP DAAC) at https://doi.org/10.5067/MEaSUREs/GFSAD/GFSAD30EUCEARUMECE.001, while the training and reference data as well as visualization are available at the Global Croplands website.

ISPRS Journal of Photogrammetry and Remote Sensing↗
USGS2021

A customized framework for regional classification of conifers using automated feature extraction

Pinyon and juniper expansion into sagebrush ecosystems is one of the major challenges facing land managers in the Great Basin. Effective pinyon and juniper treatment requires maps that accurately and precisely depict tree location and degree of woodland development so managers can target restoration efforts for early stages of pinyon and juniper expansion. However, available remotely sensed layers that cover a regional spatial extent lack the spatial resolution or accuracy to meet this need. Accuracy can be improved using object-based image analysis methods such as automated feature extraction, which has proven successful in accurately classifying land cover at the site-level but to date has yet to be applied to regional extents due to time and computational limitations. Using Feature Analyst™, we implement our framework with 1-m 2 reference imagery provided by National Agricultural Imagery Program to classify conifers across Nevada and northeastern California. Our resulting binary conifer map has an overall accuracy of 86%. We discuss the advantages to accuracy and precision our framework provides compared to other classification methods

Oregon, California, Nevada, Utah↗
USGS2007

A comparison of protocols and observer precision for measuring physical stream attributes

Stream monitoring programs commonly measure physical attributes to assess the effect of land management on stream habitat. Variability associated with the measurement of these attributes has been linked to a number of factors, but few studies have evaluated variability due to differences in protocols. We compared six protocols, five used by the U.S. Department of Agriculture Forest Service and one by the U.S. Environmental Protection Agency, on six streams in Oregon and Idaho to determine whether differences in protocol affect values for 10 physical stream attributes. Results from Oregon and Idaho were combined for groups participating in both states, with significant differences in attribute means for 9 out of the 10 stream attributes. Significant differences occurred in 5 of 10 in Idaho, and 10 of 10 in Oregon. Coefficients of variation, signal-to-noise ratio, and root mean square error were used to evaluate measurement precision. There were differences among protocols for all attributes when states were analyzed separately and as a combined dataset. Measurement differences were influenced by choice of instruments, measurement method, measurement location, attribute definitions, and training approach. Comparison of data gathered by observers using different protocols will be difficult unless a core set of protocols for commonly measured stream attributes can be standardized among monitoring programs.

Idaho, Oregon↗
USGS2002

A comparison of models for estimating the riverine export of nitrogen from large watersheds

We evaluated the accuracy of six watershed models of nitrogen export in streams (kg km2 yr-1) developed for use in large watersheds and representing various empirical and quasi-empirical approaches described in the literature. These models differ in their methods of calibration and have varying levels of spatial resolution and process complexity, which potentially affect the accuracy (bias and precision) of the model predictions of nitrogen export and source contributions to export. Using stream monitoring data and detailed estimates of the natural and cultural sources of nitrogen for 16 watersheds in the northeastern United States (drainage sizes = 475 to 70,000 km2), we assessed the accuracy of the model predictions of total nitrogen and nitrate-nitrogen export. The model validation included the use of an error modeling technique to identify biases caused by model deficiencies in quantifying nitrogen sources and biogeochemical processes affecting the transport of nitrogen in watersheds. Most models predicted stream nitrogen export to within 50% of the measured export in a majority of the watersheds. Prediction errors were negatively correlated with cultivated land area, indicating that the watershed models tended to over predict export in less agricultural and more forested watersheds and under predict in more agricultural basins. The magnitude of these biases differed appreciably among the models. Those models having more detailed descriptions of nitrogen sources, land and water attenuation of nitrogen, and water flow paths were found to have considerably lower bias and higher precision in their predictions of nitrogen export.

Biogeochemistry↗
USGS2024

A framework for disaggregating remote-sensing cropland into rainfed and irrigated classes at continental scale

Agriculture consumes the largest share of freshwater globally; therefore, distinguishing between rainfed and irrigated croplands is essential for agricultural water management and food security. In this study, a framework incorporating the Budyko model was used to differentiate between rainfed and irrigated cropland areas in Africa for eight remote sensing landcover products and a high-confidence cropland map (HCCM). The HCCM was generated for calibration and validation of the crop partitioning framework as an alternative to individual cropland masks which exhibit high disagreement. The accuracy of the framework in partitioning the HCCM was evaluated using an independent validation dataset, yielding an overall accuracy rate of 73 %. The findings of this study indicate that out of the total area covered by the HCCM (2.36 million km 2 ), about 461,000 km 2 (19 %) is irrigated cropland. The partitioning framework was applied on eight landcover products, and the extent of irrigated areas varied between 19 % and 30 % of the total cropland area. The framework demonstrated high precision and specificity scores, indicating its effectiveness in correctly identifying irrigated areas while minimizing the misclassification of rainfed areas as irrigated. This study provides an enhanced understanding of rainfed and irrigation patterns across Africa, supporting efforts towards achieving sustainable and resilient agricultural systems. Consequently, the approach outlined expands on the suite of remote sensing landcover products that can be used for agricultural water studies in Africa by enabling the extraction of irrigated and rainfed cropland data from landcover products that do not have disaggregated cropland classes.

International Journal of Applied Earth Observation↗
USGS2025

Mapping bedrock outcrops in the Sierra Nevada Mountains (California, USA) using machine learning

Accurate, high-resolution maps of bedrock outcrops can be valuable for applications such as models of land–atmosphere interactions, mineral assessments, ecosystem mapping, and hazard mapping. The increasing availability of high-resolution imagery can be coupled with machine learning techniques to improve regional bedrock outcrop maps. In the United States, the existing 30 m U.S. Geological Survey (USGS) National Land Cover Database (NLCD) tends to misestimate extents of barren land, which includes bedrock outcrops. This impacts many calculations beyond bedrock mapping, including soil carbon storage, hydrologic modeling, and erosion susceptibility. Here, we tested if a machine learning (ML) model could more accurately map exposed bedrock than NLCD across the entire Sierra Nevada Mountains (California, USA). The ML model was trained to identify pixels that are likely bedrock from 0.6 m imagery from the National Agriculture Imagery Program (NAIP). First, we labeled exposed bedrock at twenty sites covering more than 83 km 2 (0.13%) of the Sierra Nevada region. These labels were then used to train and test the model, which gave 83% precision and 78% recall, with a 90% overall accuracy of correctly predicting bedrock. We used the trained model to map bedrock outcrops across the entire Sierra Nevada region and compared the ML map with the NLCD map. At the twenty labeled sites, we found the NLCD barren land class, even though it includes more than just bedrock outcrops, accounted for only 41% and 40% of mapped bedrock from our labels and ML predictions, respectively. This substantial difference illustrates that ML bedrock models can have a role in improving land-cover maps, like NLCD, for a range of science applications.

California↗
USGS2022

Maize yield forecasts for Sub-Saharan Africa using Earth Observation data and machine learning

Food insecurity continues to grow in Sub-Saharan Africa (SSA). In 2019, chronically malnourished people numbered nearly 240 million, or 20% of the population in SSA. Globally, numerous efforts have been made to anticipate potential droughts, crop conditions, and food shortages in order to improve early warning and risk management for food insecurity. To support this goal, we develop an Earth Observation (EO) and machine-learning-based operational, subnational maize yield forecast system and evaluate its out-of-sample forecast skills during the growing seasons for Kenya, Somalia, Malawi, and Burkina Faso. In general, forecast skills improve substantially during the vegetative growth period (VP) and gradually during the reproductive development period (RP). Thus, mid-season assessment can provide effective early warning months before harvest. Skillful forecasts (Nash Sutcliffe Efficiency (NSE) > 0.6 and Mean Absolute Percentage Error (MAPE) < 20%) appear approximately two dekads after the VP; for example, skillful forecasts appear in May in Kenya and Somalia, January in Malawi, and July in Burkina Faso. During model development, effective EO features are also identified, such as precipitation and available water during VP, and dry days and extreme temperatures in early VP. Compared to monthly standard EO features, sub-monthly (dekadal), non-standard, and serial EO features significantly improve forecast skills by ＋ 0.3 NSE and -10% of MAPE, demonstrating the ability to precisely and effectively capture favorable or detrimental crop development conditions. Finally, skillful forecasts and practical utility are demonstrated in the recent normal and dry years in each region. Overall, the developed yield forecasting system can provide skillful predictions during the growing season, supporting regional and international agricultural decision-making processes, including informing food-security planning and management, thereby helping to mitigate food shortages caused by unfavorable climate conditions.

Global Food Security↗
USGS2009

Concentrations, and estimated loads and yields of total nitrogen and total phosphorus at selected water-quality monitoring network stations in Kentucky, 1979-2004

To evaluate the State’s water quality, the Kentucky Division of Water collects data from a statewide network of primary ambient stream water-quality monitoring stations and flexible, rotating watershed-monitoring stations. This ambient stream water-quality monitoring network program is directed to assess the conditions of surface waters throughout Kentucky. Water samples were collected monthly for the majority of the stations from 1979 to 1998, which represented agricultural, undeveloped (mainly forested), and areas of mixed land use/land cover. In 1998, the number of water samples collected was reduced to a collection frequency of six times per year (every 2 months) every 4 of 5 years, because a new monitoring network was implemented involving a 5-year rotating Basin Management Unit scheme of monitoring. This report presents the results of a study conducted by the U.S. Geological Survey, in cooperation with the Kentucky Energy and Environment Cabinet–Kentucky Division of Water, to summarize concentrations of total nitrogen and total phosphorus and provide estimates of total nitrogen and total phosphorus loads and yields in 55 selected streams in Kentucky’s ambient stream water-quality monitoring network, which was operated from 1979 through 2004. Streams in predominately agricultural basins had higher concentrations of total nitrogen (TN) and concentrations of total phosphorus (TP) than streams in predominately undeveloped (forested) basins. Streams in basins in intensely developed karst areas characterized by caves, springs, sinkholes, and sinking streams had a higher median concentration of TN (1.5 milligrams per liter [mg/L]) than streams in basins with limited or no karst areas (0.63 mg/L). As with TN, median concentrations of TP also were higher in areas of intense karst (0.05 mg/L) than in areas with limited or no karst (0.02 mg/L). The U.S. Environmental Protection Agency (USEPA) has recommended ecoregional nutrient water-quality criteria as a starting point for States to establish more precise numeric water-quality criteria for nutrients to protect aquatic life and recreational and other uses of rivers and streams. On the basis of the 25 th percentile of concentration data from reference stations aggregated by ecoregion, the USEPA established recommended water-quality criteria for TN and TP in the two Aggregated Ecoregions (IX and XI) in Kentucky waters. The 25 th percentile median values for TN and TP from this study exceeded the USEPA’s recommendations in both aggregated ecoregions in the agricultural and mixed land-use/land-cover basins, and for TN in the undeveloped land-use/land-cover basins in Aggregated Ecoregion XI. However, the 25 th percentile median values for TN (Aggregated Ecoregion IX) and TP in both aggregated ecoregions did not exceed the USEPA’s recommendations in the undeveloped land-use/land-cover basins. Estimated loads and yields of TN and TP varied substantially among the individual stations. Estimated mean annual yields of TN ranged from 0.10 [tons per year per square mile (ton/yr)/mi 2 ] to 7.2 (ton/yr)/mi 2 , and estimated mean annual yields of TP ranged from 0.02 (ton/yr)/mi 2 to 1.4 (ton/yr)/mi 2 . Estimated mean annual yields of TN and TP were generally highest at stations in predominately agricultural basins, and lowest at stations in undeveloped land-use/land-cover basins.

Kentucky↗
USGS2020

Molecular sequencing and morphological identification reveal similar patterns in native bee communities across public and private grasslands of eastern North Dakota

Bees play a key role in the functioning of human-modified and natural ecosystems by pollinating agricultural crops and wild plant communities. Global pollinator conservation efforts need large-scale and long-term monitoring to detect changes in species’ demographic patterns and shifts in bee community structure. The objective of this project was to test a molecular sequencing pipeline that would utilize a commonly used locus, produce accurate and precise identifications consistent with morphological identifications, and generate data that are both qualitative and quantitative. We applied this amplicon sequencing pipeline to native bee communities sampled across Conservation Reserve Program (CRP) lands and native grasslands in eastern North Dakota. We found the 28S LSU locus to be more capable of discriminating between species than the 18S SSU rRNA locus, and in some cases even resolved instances of cryptic species or morphologically ambiguous species complexes. Overall, we found the amplicon sequencing method to be a qualitatively accurate representation of the sampled bee community richness and species identity, especially when a well-curated database of known 28S LSU sequences is available. Both morphological identification and molecular sequencing revealed similar patterns in native bee community structure across CRP lands and native prairie. Additionally, a genetic algorithm approach to compute taxon-specific correction factors using a small subset of the most concordant samples demonstrated that a high level of quantitative accuracy could be possible if the specimens are fresh and processed soon after collection. Here we provide a first step to a molecular pipeline for identifying insect pollinator communities. This tool should prove useful for future national monitoring efforts as use of molecular tools becomes more affordable and as numbers of 28S LSU sequences for pollinator species increase in publicly-available databases.

North Dakota↗
USGS2026

Assessing the state of hydrologic science in the Upper Klamath Basin—A comprehensive review of data, tools, and models

Water demand in the Upper Klamath Basin (UKB) from various stakeholders and ecological needs often outstrips available supply, leading to persistent management challenges. This study reviews the state of hydrologic science within the UKB as of 2025—specifically, the tools, data, and models available for assessing five key components of the water system: (1) surface water; (2) precipitation; (3) evapotranspiration; (4) groundwater; and (5) water use. The UKB water supply is critical for Native American communities, regional agriculture, and federally listed fishes and faces challenges from competing needs, climate variability, and operational/regulatory requirements. We assess existing datasets, regional and national models, and historical studies to understand the available resources and identify gaps that may hinder integrated water assessments and management. Our findings indicate areas where improvements in data collection and model precision could improve the accuracy of water-availability forecasts and support water-management practices. This review can inform near-term forecasting, assist in optimizing water-resource data collection and management strategies, and support regional water-availability assessments of the basin.

Oregon↗
USGS2023

Groundwater-flow model of the Treasure Valley, southwestern Idaho, 1986–2015

Most of the population of the Treasure Valley and the surrounding area of southwestern Idaho and easternmost Oregon depends on groundwater for domestic supply, either from domestic or municipal-supply wells. Current and projected rapid population growth in the area has caused concern about the long-term sustainability of the groundwater resource. In 2016, the U.S. Geological Survey, in cooperation with the Idaho Water Resource Board and the Idaho Department of Water Resources, began a project to construct a numerical groundwater-flow model of the westernmost portion of the western Snake River Plain aquifer system, called the Treasure Valley. The development of the model was guided by several objectives, including: to improve the understanding of groundwater and surface water interactions; to facilitate conjunctive water management; to provide a tool for water resources planning; and to provide a tool for water allocation. The model was constructed with a spatial scale and level of detail that aimed to meet these objectives while balancing the sometimes-competing goals of fast runtimes, numerical stability, usability, and parsimony. The Treasure Valley Groundwater Flow Model (TVGWFM) is a three-dimensional finite-difference numerical model constructed using MODFLOW 6 (Langevin and others, 2017, Documentation for the MODFLOW 6 Groundwater Flow Model: U.S. Geological Survey Techniques and Methods, book 6, chap. A55, 197 p., https://doi.org/10.3133/tm6A55 ). The model covers the westernmost portion of the western Snake River Plain and is discretized into a regular grid of 64 by 65 cells with a side length of 1 mile and 6 layers of varying depth and active area. A historical model period was developed consisting of 360 month-long stress periods for 1986–2015. The model builds upon previous modeling efforts by adding a transient period, incorporating new head and discharge observations to constrain parameters, incorporating information from the hydrogeologic framework model (HFM) of Bartolino (2019, Hydrogeologic framework of the Treasure Valley and surrounding area, Idaho and Oregon: U.S. Geological Survey Scientific Investigations Report 2019–5138, https://doi.org/10.3133/sir20195138 ) and incorporating refined estimates of evapotranspiration and irrigation classification of lands in the study area. The TVGWFM includes all significant components of recharge to and discharge from the aquifer. Inflows include canal seepage, irrigation and precipitation recharge, mountain-front recharge, rivers and stream seepage, and seepage from Lake Lowell. Outflows include discharge to agricultural drainage ditches, discharge to rivers and streams, pumping, and discharge to Lake Lowell. Each recharge or discharge component is represented separately using individual MODFLOW 6 packages. Parameter values were derived with a combination of trial-and-error steps and automated parameter estimation using PEST software (Doherty, J.E., 2005, PEST, model-independent parameter estimation–User manual: Watermark Numerical Computing, https://pesthomepage.org/documentation ). Parameter estimates were constrained with several types of observation data, including water levels, water level changes, vertical water level differences, drain discharges, change in drain discharges, river seepage, seepage from Lake Lowell, and change in seepage from Lake Lowell. Material properties from the hydrogeologic framework were also used to assign the minimum and maximum values of some parameters. A final parameter realization was reached that minimized residuals between the observed and modelled values for the various observation groups. Mean residuals for the observation groups were 15.4 feet (ft) for water levels, 0.2 ft for water level changes, 19.4 ft for vertical water level differences, −3.9 cubic feet per second (ft 3 /s) for drain discharges, 0.0 ft 3 /s for changes in drain discharge, 45.0 ft 3 /s for river seepage, −40.1 ft 3 /s for Lake Lowell seepage, and 126.3 ft 3 /s for changes in Lake Lowell seepage. The quality of the model’s fit to observations varied spatially, with notable areas of under- or over-simulation of water levels present to the northwest and southwest of Lake Lowell, in the foothills along the eastern model boundary, and near the City of Eagle. Trends were observed in the residuals of many of the observation groups, indicating that the model is missing or not fully reproducing some phenomena that are observed in the system. The TVGWFM can be used as a tool for water resource planning, for understanding the interactions of groundwater and surface water at a basin scale, and for facilitating conjunctive management, but may lack the precision needed for water rights administration at a local scale. Additional sources of uncertainty or limitations of the model are noted. The quantity and spatial distribution of canal seepage and infiltration of irrigation water recharge, the largest sources of recharge to the system, are unknown and approximated indirectly. There is poor understanding of how canal seepage and incidental recharge change as land is converted from agricultural (irrigated) to suburban (semi-irrigated). These uncertainties will affect any scenarios that investigate changes to land use or irrigation practices. Finally, the model has relatively high water-level residuals around and to the southwest of Lake Lowell and should not be used to estimate water level effects in that region. The model was built with multiple, broadly expressed objectives and did not optimize performance for specific uses. However, the model and the tools included in an associated data release provide ample flexibility to improve the model for future uses. Adjustments and improvements could be made by refining the model in an area of interest, collecting additional calibration data, applying more rigorous boundary conditions, or re-estimating model parameters to optimize model performance for a specific model forecast.

Idaho↗