Russell G. Congalton

Determining Forest species composition using high spectral resolution remote sensing data

Airborne hyperspectral data were analyzed for the classification of 11 forest cover types, including pure and mixed stands of deciduous and conifer species. Selected bands from first difference reflectance spectra were used to determine cover type at the Harvard Forest using a maximum likelihood algorithm assigning all pixels in the image into one of the 11 categories. This approach combines species specific chemical characteristics and previously derived relationships between hyperspectral data and foliar chemistry. Field data utilized for validation of the classification included both a stand-level survey of stem diameter, and field measurements of plot level foliar biomass. A random selection of validation pixels yielded an overall classification accuracy of 75%.

A Comparison of Urban Mapping Methods Using High-Resolution Digital Imagery

Recent advances in digital airborne sensors and satellite platforms make spatially accurate, high-resolution multispectral imagery readily available. These advances provide the opportunity for a host of new applications to address and solve old problems. High-resolution imagery is particularly well suited to urban applications. Previous data sources (such as Landsat TM) did not show the spatial detail necessary to provide many urban planning solutions. This paper provides an overview of a project in which one-meter digital imagery was used to produce a map of pervious and impervious surfaces to be used by the city of Scottsdale, Arizona for storm-water runoff estimation. The increased spatial information in onemeter or less resolution imagery strains the usefulness of image classification using traditional supervised and unsupervised spectral classification algorithms. This study assesses the accuracy of three different methods for extracting land-cover/land-use information from high-resolution imagery of urban environments: (1) combined supervised/ unsupervised spectral classification, (2) raster-based spatial modeling, and (3) image segmentation classification using classification tree analysis. A discussion of the results and relative merits of each method is included.

Evaluating the potential for measuring river discharge from space

Numerous studies have demonstrated the potential usefulness of river hydraulic data obtained from satellites in developing general approaches to tracking floods and changes in river discharge from space. Few studies, however, have attempted to estimate the magnitude of discharge in rivers entirely from remotely obtained information. The present study uses multiple-regression analyses of hydraulic data from more than 1000 discharge measurements, ranging in magnitude from over 200,000 to less than 1 m3/s, to develop multi-variate river discharge estimating equations that use various combinations of potentially observable variables to estimate river discharge. Uncertainty analysis indicates that existing satellite-based sensors can measure water-surface width (or surface area), water-surface elevation, and potentially the surface velocity of rivers with accuracies sufficient to provide estimates of discharge with average uncertainty of less than 20%. Development and validation of multi-variate rating equations that are applicable to the full range of rivers that can be observed from satellite sensors, development of techniques to accurately estimate the average depth in rivers from stage measurements, and development of techniques to accurately estimate the average velocity in rivers from surface-velocity measurements will be key to successful prediction of discharge from satellite observations.

Evaluating remotely sensed techniques for mapping riparian vegetation

Riparian vegetation provides many important functions for fish and wildlife habitat, yet we lack the information about riparian condition needed for planning and policy decisions. In this project, we developed a more cost-effective method to classify riparian vegetation using aerial photography and compared the results with maps generated from classifying Landsat Thematic Mapper (TM) imagery. We used ortho-photography and other data freely available from the public domain, plus ArcInfos capability to display stream buffers on top of photos, to identify the vegetation structure and location. Two buffers, one from 0 to 15.25 m and the other from 15.25 to 61 m on each side of the streams, were segmented rather than using the traditional method of drawing polygons around the vegetation types. This method was faster and cheaper than digitizing vegetation polygons, plus it allowed the photo interpreters judgment to correct for stream location errors in the data without having to redraw the streams location. Riparian vegetation structure was found to be far different from upland forest stands in the watershed. Hardwoods dominated the riparian zones. Overall, hardwood stands made up 59% of the vegetation within the first 15.25 m from the stream in a watershed where 75% of the upland area was conifer stands. Agricultural zones were dominated by hardwood stands and open conditions. Only 1–2% of the areas within 15.25 m of the streams had large conifer stands, while over 80% of these areas were hardwoods or brush. The ‘open area’ class comprised 13% of the total area within the first 15.25 m of the streams and increased to 49% in the area between 15.25 and 61 m from the streams. Current forest classifications based on Landsat TM imagery did not do a good job of identifying the structural characteristics of riparian vegetation. Our Landsat TM and photo classifications only agreed 25–30% of the time. The extreme diversity and linear arrangement of the riparian vegetation creates classification problems and results in the Landsat TM imagery being inadequate for use in policy decisions.

Global Land Cover Mapping: A Review and Uncertainty Analysis

Given the advances in remotely sensed imagery and associated technologies, several global land cover maps have been produced in recent times including IGBP DISCover, UMD Land Cover, Global Land Cover 2000 and GlobCover 2009. However, the utility of these maps for specific applications has often been hampered due to considerable amounts of uncertainties and inconsistencies. A thorough review of these global land cover projects including evaluating the sources of error and uncertainty is prudent and enlightening. Therefore, this paper describes our work in which we compared, summarized and conducted an uncertainty analysis of the four global land cover mapping projects using an error budget approach. The results showed that the classification scheme and the validation methodology had the highest error contribution and implementation priority. A comparison of the classification schemes showed that there are many inconsistencies between the definitions of the map classes. This is especially true for the mixed type classes for which thresholds vary for the attributes/discriminators used in the classification process. Examination of these four global mapping projects provided quite a few important lessons for the future global mapping projects including the need for clear and uniform definitions of the classification scheme and an efficient, practical, and valid design of the accuracy assessment.

Sampling method and sample placement: How do they affect the accuracy of remotely sensed maps?

for accuracy against the reference data. A widely accepted proThe accuracy of remotely sensed forest stand maps is tra- cedure for comparing these data is the generation of an error ditionally assessed by comparing a sample of the map data matrix (Card, 1982; Congalton et al., 1983; Story and Congalton, with actual ground conditions. Samples most often comprise 1986; Congalton, 1991). clusters of pixels within homogeneous areas, thereby avoiding An error matrix is an especially effective accuracy assessproblems associated with accurately mapping “edges” (e.g., ment tool because it provides a starting point for a series of statransition areas between two forest types). Consequently, they tistical techniques to further examine accuracy (Congalton and may well overestimate accuracy, but the degree of overestima- Green, 1999). One such analytical technique is the Kappa analtion is unknown. This paper examines two important factors ysis, a discrete multivariate technique for comparing error maregarding accuracy assessment that are not well studied: the trices (Congalton et al., 1983; Hudson and Ramm, 1987; Coneffect on estimates of accuracy of (1) the sampling method galton, 1991; Ma and Redmond, 1995; Stehman, 1996; Stehand (2) the exact placement of the samples. Overall accuracy, man, 1999; Congalton and Green, 1999). Kappa analysis, which normalized accuracy, and the KHAT statistic are computed from assumes a multinomial distribution, generates a KHAT statistic error matrices generated from simple random sampling, stra- that measures the difference between actual and chance (or rantified random sampling, and systematic sampling using totally dom) agreement between the map and reference data. It can also random sample placement and samples chosen from homog- be used to test for significant differences between two error eneous areas only. The results indicate that Kappa appears matrices. to be as appropriate to use with systematic sampling and The only sampling method that satisfies Kappa’s assumpstratified random sampling as it is with simple random sam- tion of a multinomial model is simple random sampling. The pling, but suggests that sample placement may have more of effect of other sampling schemes on the outcome of the Kappa an effect on estimates of accuracy than sampling method analysis has not been well studied. In addition, samples are ofalone. ten chosen only if they occur within the interior of homogeneous pixel groupings in order to avoid problems with sam

Nominal 30-M Cropland Extent Map of Continental Africa by Integrating Pixel-Based and Object-Based Algorithms Using Sentinel-2 and Landsat-8 Data on Google Earth Engine

A satellite-derived cropland extent map at high spatial resolution (30-m or better) is a must for food and water security analysis. Precise and accurate global cropland extent maps, indicating cropland and non-cropland areas, are starting points to develop higher-level products such as crop watering methods (irrigated or rainfed), cropping intensities (e.g., single, double, or continuous cropping), crop types, cropland fallows, as well as for assessment of cropland productivity (productivity per unit of land), and crop water productivity (productivity per unit of water). Uncertainties associated with the cropland extent map have cascading effects on all higher-level cropland products. However, precise and accurate cropland extent maps at high spatial resolution over large areas (e.g., continents or the globe) are challenging to produce due to the small-holder dominant agricultural systems like those found in most of Africa and Asia. Cloud-based geospatial computing platforms and multi-date, multi-sensor satellite image inventories on Google Earth Engine offer opportunities for mapping croplands with precision and accuracy over large areas that satisfy the requirements of broad range of applications. Such maps are expected to provide highly significant improvements compared to existing products, which tend to be coarser in resolution, and often fail to capture fragmented small-holder farms especially in regions with high dynamic change within and across years. To overcome these limitations, in this research we present an approach for cropland extent mapping at high spatial resolution (30-m or better) using the 10-day, 10 to 20-m, Sentinel-2 data in combination with 16-day, 30-m, Landsat-8 data on Google Earth Engine (GEE). First, nominal 30-m resolution satellite imagery composites were created from 36,924 scenes of Sentinel-2 and Landsat-8 images for the entire African continent in 2015–2016. These composites were generated using a median-mosaic of five bands (blue, green, red, near-infrared, NDVI) during each of the two periods (period 1: January–June 2016 and period 2: July–December 2015) plus a 30-m slope layer derived from the Shuttle Radar Topographic Mission (SRTM) elevation dataset. Second, we selected Cropland/Non-cropland training samples (sample size = 9791) from various sources in GEE to create pixel-based classifications. As supervised classification algorithm, Random Forest (RF) was used as the primary classifier because of its efficiency, and when over-fitting issues of RF happened due to the noise of input training data, Support Vector Machine (SVM) was applied to compensate for such defects in specific areas. Third, the Recursive Hierarchical Segmentation (RHSeg) algorithm was employed to generate an object-oriented segmentation layer based on spectral and spatial properties from the same input data. This layer was merged with the pixel-based classification to improve segmentation accuracy. Accuracies of the merged 30-m crop extent product were computed using an error matrix approach in which 1754 independent validation samples were used. In addition, a comparison was performed with other available cropland maps as well as with LULC maps to show spatial similarity. Finally, the cropland area results derived from the map were compared with UN FAO statistics. The independent accuracy assessment showed a weighted overall accuracy of 94%, with a producer’s accuracy of 85.9% (or omission error of 14.1%), and user’s accuracy of 68.5% (commission error of 31.5%) for the cropland class. The total net cropland area (TNCA) of Africa was estimated as 313 Mha for the nominal year 2015. The online product, referred to as the Global Food Security-support Analysis Data @ 30-m for the African Continent, Cropland Extent product (GFSAD30AFCE) is distributed through the NASA’s Land Processes Distributed Active Archive Center (LP DAAC) as (available for download by 10 November 2017 or earlier): https://doi.org/10.5067/MEaSUREs/GFSAD/GFSAD30AFCE.001 and can be viewed at https://croplands.org/app/map. Causes of uncertainty and limitations within the crop extent product are discussed in detail.

Tools for Successful Student-Teacher-Scientist Partnerships.

Student–teacher–scientist partnership (STSP) programs are cooperative relationships in which students, with the support of their teachers, participate in and contribute to the research of scientists. This paper examines one of the worlds largest STSPs—an international environmental science education program called GLOBE (Global Learning and Observations to Benefit the Environment)—and proposes recommendations to scientists about how they can get the most out of their research and teaching relationship with students and their teachers. GLOBE is an international K–12 STSP that engages students in Earths Systems investigations. Extensive training is needed for students to collect and report accurate data to scientists, and special preparatory curricula are needed to make their partnership effective and motivating. Recognizing these issues, this research was conducted specifically to identify and recommend a set of training material design criteria for implementation of STSPs in the elementary and middle school levels. The conclusions—the result of background research, extensive interviews and consultation with teachers—provide guidance to GLOBE and other STSP programs to enhance the development of effective and engaging training materials.

A GLOBE Collaboration to Develop Land Cover Data Collection and Analysis Protocols

Global Learning and Observations to Benefit the Environment (GLOBE) is an international environmental education and science partnership which coordinates the work of students (aged 5 to 18), teachers and scientists from 48 countries on five continents to study and better understand the global environment. Accurate ground reference data is fundamental to the use of remotely sensed data for land cover classification and mapping. Because very little ground reference data has been collected, the accuracy of many land cover maps may be questioned, thus accurate land cover ground reference data is an important need that could be addressed through GLOBE scientist-student collaboration. If earth systems scientists are to use student data, it is important that those data be as accurate as possible to ensure reliability of research results. Thus a key question for this research is whether student collected data are accurate enough to support rigorous scientific investigations. This paper describes results of the GLOBE Science-Education Team on Data Validation and Accuracy Assessments collaboration with teachers and students to: (1) design and test the pre-protocol learning activities; (2) test the protocols intended to guide the collection and analysis of data; and (3) implement the learning activities and protocols to determine the relative accuracy of student collected versus professionally collected land cover data. To ensure the most accurate classification of land cover possible, a new international hierarchical land cover classification system, the Modified Unesco Classification (MUC) system was developed. GLOBE Data Collection Protocols and methods were designed and implemented to test the accuracy of student collected reference data were designed and implemented. Students who collected land cover reference data using GLOBE protocols, obtained data which are at least as accurate as that collected by professionals.

Spectral matching techniques (SMTs) and automated cropland classification algorithms (ACCAs) for mapping croplands of Australia using MODIS 250-m time-series (2000–2015) data

ABSTRACT Mapping croplands, including fallow areas, are an important measure to determine the quantity of food that is produced, where they are produced, and when they are produced (e.g. seasonality). Furthermore, croplands are known as water guzzlers by consuming anywhere between 70% and 90% of all human water use globally. Given these facts and the increase in global population to nearly 10 billion by the year 2050, the need for routine, rapid, and automated cropland mapping year-after-year and/or season-after-season is of great importance. The overarching goal of this study was to generate standard and routine cropland products, year-after-year, over very large areas through the use of two novel methods: (a) quantitative spectral matching techniques (QSMTs) applied at continental level and (b) rule-based Automated Cropland Classification Algorithm (ACCA) with the ability to hind-cast, now-cast, and future-cast. Australia was chosen for the study given its extensive croplands, rich history of agriculture, and yet nonexistent routine yearly generated cropland products using multi-temporal remote sensing. This research produced three distinct cropland products using Moderate Resolution Imaging Spectroradiometer (MODIS) 250-m normalized difference vegetation index 16-day composite time-series data for 16 years: 2000 through 2015. The products consisted of: (1) cropland extent/areas versus cropland fallow areas, (2) irrigated versus rainfed croplands, and (3) cropping intensities: single, double, and continuous cropping. An accurate reference cropland product (RCP) for the year 2014 (RCP2014) produced using QSMT was used as a knowledge base to train and develop the ACCA algorithm that was then applied to the MODIS time-series data for the years 2000–2015. A comparison between the ACCA-derived cropland products (ACPs) for the year 2014 (ACP2014) versus RCP2014 provided an overall agreement of 89.4% (kappa = 0.814) with six classes: (a) producer’s accuracies varying between 72% and 90% and (b) user’s accuracies varying between 79% and 90%. ACPs for the individual years 2000–2013 and 2015 (ACP2000–ACP2013, ACP2015) showed very strong similarities with several other studies. The extent and vigor of the Australian croplands versus cropland fallows were accurately captured by the ACCA algorithm for the years 2000–2015, thus highlighting the value of the study in food security analysis. The ACCA algorithm and the cropland products are released through http://croplands.org/app/map and http://geography.wr.usgs.gov/science/croplands/algorithms/australia_250m.html