Search

Although rice cultivation is one of the most important agricultural sources of methane (CH4) and contributes ∼8% of total global anthropogenic emissions, large discrepancies remain among estimates of global CH4 emissions from rice cultivation (ranging from 18 to 115 Tg CH4 yr−1) due to a lack of observational constraints. The spatial distribution of paddy-rice emissions has been assessed at regional-to-global scales by bottom-up inventories and land surface models over coarse spatial resolution (e.g., > 0.5°) or spatial units (e.g., agro-ecological zones). However, high-resolution CH4 flux estimates capable of capturing the effects of local climate and management practices on emissions, as well as replicating in situ data, remain challenging to produce because of the scarcity of high-resolution maps of paddy-rice and insufficient understanding of CH4 predictors. Here, we combine paddy-rice methane-flux data from 23 global eddy covariance sites and MODIS remote sensing data with machine learning to 1) evaluate data-driven model performance and variable importance for predicting rice CH4 fluxes; and 2) produce gridded up-scaling estimates of rice CH4 emissions at 5000-m resolution across Monsoon Asia, where ∼87% of global rice area is cultivated and ∼ 90% of global rice production occurs. Our random-forest model achieved Nash-Sutcliffe Efficiency values of 0.59 and 0.69 for 8-day CH4 fluxes and site mean CH4 fluxes respectively, with land surface temperature, biomass and water-availability-related indices as the most important predictors. We estimate the average annual (winter fallow season excluded) paddy rice CH4 emissions throughout Monsoon Asia to be 20.6 ± 1.1 Tg yr−1 for 2001–2015, which is at the lower range of previous inventory-based estimates (20–32 CH4 Tg yr−1). Our estimates also suggest that CH4 emissions from paddy rice in this region have been declining from 2007 through 2015 following declines in both paddy-rice growing area and emission rates per unit area, suggesting that CH4 emissions from paddy rice in Monsoon Asia have likely not contributed to the renewed growth of atmospheric CH4 in recent years.

Tidal flats (non-vegetated area), along with coastal vegetation area, constitute the coastal wetlands (intertidal zone) between high and low water lines, and play an important role in wildlife, biodiversity and biogeochemical cycles. However, accurate annual maps of coastal tidal flats over the last few decades are unavailable and their spatio-temporal changes in China are unknown. In this study, we analyzed all the available Landsat TM/ETM+/OLI imagery (~44,528 images) using the Google Earth Engine (GEE) cloud computing platform and a robust decision tree algorithm to generate annual frequency maps of open surface water body and vegetation to produce annual maps of coastal tidal flats in eastern China from 1986 to 2016 at 30-m spatial resolution. The resulting map of coastal tidal flats in 2016 was evaluated using very high-resolution images available in Google Earth. The total area of coastal tidal flats in China in 2016 was about 731,170 ha, mostly distributed in the provinces around Yellow River Delta and Pearl River Delta. The interannual dynamics of coastal tidal flats area in China over the last three decades can be divided into three periods: a stable period during 1986–1992, an increasing period during 1993–2001 and a decreasing period during 2002–2016. The resulting annual coastal tidal flats maps could be used to support sustainable coastal zone management policies that preserve coastal ecosystem services and biodiversity in China.

Over the past few decades, wide encroachment of eastern redcedar (Juniperus virginiana) and Ashe juniper (Juniperus ashei) into the prairies of the U.S. Great Plains has affected wildlife habitats, forage and livestock production, and biogeochemical cycles. This study investigates the spatio-temporal dynamics of juniper forest encroachment into tallgrass prairies by generating juniper forest encroachment maps from 1984 to 2010 at 30 m spatial resolution. A pixel and phenology-based mapping algorithm was used to produce the time series maps of juniper forest encroachment using a combination of Phased Array type L-band Synthetic Aperture Radar (PALSAR) mosaic data from 2010 and Landsat 5 and 7 data (10,871 images from 1984 to 2010). We analyzed the resultant maps to understand the dynamics of juniper forest encroachment at state and county spatial scales and examined juniper occurrence by geographic region and soil type. The juniper forest maps were generated over five multi-year periods: the late 1980s (1984–1989), early 1990s (1990–1994), late 1990s (1995–1999), early 2000s (2000–2004), and late 2000s (2005–2010). We also produced a map of time since stand detection of juniper forests in 2010. Our major findings include: (1) juniper forests have expanded linearly in time at an annual rate of ~ 40 km2/year since 1984; (2) juniper forests had notable spatial clusters in its expansion process; (3) ~ 65% of juniper forests in 2010 were

South America has the largest tropical rainforests and the richest biodiversity in the world. It is challenging to map tropical forests and their spatio-temporal changes because forests are facing fragmentation from human activities (e.g., logging, deforestation), drought, and fire, as well as persistent clouds. Here we present a robust approach to map forests in South America during 2007–2010 and analyze the consistency and uncertainty among eight major forest maps in South America. Greenness-relevant MOD13Q1 NDVI and structure/biomass-relevant ALOS PALSAR time series data recorded 2007 through 2010 were coupled to identify and map forests at 50-m spatial resolution. Both area and spatial comparison were conducted to analyze the consistency and uncertainty of these eight forest maps. Annual 50-m PALSAR/MODIS forest maps were generated during 2007–2010 and the total forest area in South America was about 8.63 × 10 6 km 2 in 2010. Large differences in total forest area (8.2 × 10 6 km 2 –12.7 × 10 6 km 2 ) existed among these forest products, especially in the forest edges, semi-humid tropical, and subtropical regions. Forest products generated under a similar forest definition had similar or even larger variation than those generated with contrasting forest definitions. We also find out that one needs to consider leaf area index as an adjusting factor and use much higher threshold values in the Vegetation Continuous Field (VCF) datasets to estimate forest cover areas. Analyses of PALSAR/MODIS forest maps in 2008/2009 showed a relatively small rate of loss (3.2 × 10 4 km 2 year − 1 ) in net forest cover, similar to that of FAO FRA (3.3 × 10 4 km 2 year − 1 ), but much higher annual rates of forest loss and gain. The rate of forest loss (0.195 × 10 6 km 2 year − 1 ) was higher than that of Global Forest Watch (0.081 × 10 6 km 2 year − 1 ). PALSAR/MODIS forest maps showed that more deforestation occurred in the unfragmented forest areas. Caution should be used when using the different forest maps to analyze forest loss and make policies regarding forest ecosystem services and biodiversity conservation. The integration of PALSAR and MODIS images during 2007–2010 provides annual maps of forests in South America with improved accuracy and reduced uncertainty.

Woody plant encroachment of eastern redcedar (Juniperus virginiana L., hereafter referred to as “red cedar”) into native grasslands in the U.S. Southern Great Plains has significantly affected the production of forage and livestock, wildlife habitats, as well as water, carbon, nutrient and biogeochemical cycles. However, time series of red cedar maps are still not available to document the continuously spatio-temporal dynamics of red cedar encroachment across landscape, watershed and regional scales. In this study, we developed a pixel and phenology-based mapping algorithm, and used it to analyze PALSAR mosaic data in 2010 and all the available Landsat 5/7 data during 1984–2010 with the Google Earth Engine (GEE) platform. This pilot study analyzed 4233 images covering > 10 counties in the central region of Oklahoma, and generated red cedar forest maps for 2010 and five historical time periods: the late 1980s (1984–1989), early 1990s (1990–1994), late 1990s (1995–1999), early 2000s (2000–2004), and late 2000s (2005–2010). The resultant maps for 2010, the late 2000s, early 2000s, and late 1990s were evaluated using validation samples collected from Google Earth's high-resolution images and geo-referenced field photos. The overall (producer and user) accuracy of these maps ranged from 88% to 96% (88%–93%, and 96%–99%). The resultant maps clearly illustrated an increase in red cedar encroachment within the study area at an annual rate of ~ 8% during 1984–2010. These maps can be used to support additional studies on the driving factors and consequences of red cedar encroachment. This study also demonstrated the potential to trace the historical encroachment of red cedar into grasslands using time series Landsat images and PALSAR data.

Area and spatial distribution information of paddy rice are important for understanding of food security, water use, greenhouse gas emission, and disease transmission. Due to climatic warming and increasing food demand, paddy rice has been expanding rapidly in high latitude areas in the last decade, particularly in northeastern (NE) Asia. Current knowledge about paddy rice fields in these cold regions is limited. The phenology- and pixel-based paddy rice mapping (PPPM) algorithm, which identifies the flooding signals in the rice transplanting phase, has been effectively applied in tropical areas, but has not been tested at large scale of cold regions yet. Despite the effects from more snow/ice, paddy rice mapping in high latitude areas is assumed to be more encouraging due to less clouds, lower cropping intensity, and more observations from Landsat sidelaps. Moreover, the enhanced temporal and geographic coverage from Landsat 8 provides an opportunity to acquire phenology information and map paddy rice. This study evaluated the potential of Landsat 8 images on annual paddy rice mapping in NE Asia which was dominated by single cropping system, including Japan, North Korea, South Korea, and NE China. The cloud computing approach was used to process all the available Landsat 8 imagery in 2014 (143 path/rows, ~ 3290 scenes) with the Google Earth Engine (GEE) platform. The results indicated that the Landsat 8, GEE, and improved PPPM algorithm can effectively support the yearly mapping of paddy rice in NE Asia. The resultant paddy rice map has a high accuracy with the producer (user) accuracy of 73% (92%), based on the validation using very high resolution images and intensive field photos. Geographic characteristics of paddy rice distribution were analyzed from aspects of country, elevation, latitude, and climate. The resultant 30-m paddy rice map is expected to provide unprecedented details about the area, spatial distribution, and landscape pattern of paddy rice fields in NE Asia, which will contribute to food security assessment, water resource management, estimation of greenhouse gas emissions, and disease control.

Accurate estimation of the gross primary production (GPP) of terrestrial ecosystems is vital for a better understanding of the spatial-temporal patterns of the global carbon cycle. In this study,we estimate GPP in North America (NA) using the satellite-based Vegetation Photosynthesis Model (VPM), MODIS (Moderate Resolution Imaging Spectrometer) images at 8-day temporal and 500 meter spatial resolutions, and NCEP-NARR (National Center for Environmental Prediction-North America Regional Reanalysis) climate data. The simulated GPP (GPP (sub VPM)) agrees well with the flux tower derived GPP (GPPEC) at 39 AmeriFlux sites (155 site-years). The GPP (sub VPM) in 2010 is spatially aggregated to 0.5 by 0.5-degree grid cells and then compared with sun-induced chlorophyll fluorescence (SIF) data from Global Ozone Monitoring Instrument 2 (GOME-2), which is directly related to vegetation photosynthesis. Spatial distribution and seasonal dynamics of GPP (sub VPM) and GOME-2 SIF show good consistency. At the biome scale, GPP (sub VPM) and SIF shows strong linear relationships (R (sup 2) is greater than 0.95) and small variations in regression slopes ((4.60-5.55 grams Carbon per square meter per day) divided by (milliwatts per square meter per nanometer per square radian)). The total annual GPP (sub VPM) in NA in 2010 is approximately 13.53 petagrams Carbon per year, which accounts for approximately 11.0 percent of the global terrestrial GPP and is within the range of annual GPP estimates from six other process-based and data-driven models (11.35-22.23 petagrams Carbon per year). Among the seven models, some models did not capture the spatial pattern of GOME-2 SIF data at annual scale, especially in Midwest cropland region. The results from this study demonstrate the reliable performance of VPM at the continental scale, and the potential of SIF data being used as a benchmark to compare with GPP models.

Sugarcane is a major crop for sugar and ethanol production and its area has increased substantially in tropical and subtropical regions in recent decades. Updated and accurate sugarcane maps are critical for monitoring sugarcane area and production and assessing its impacts on the society, economy and the environment. To date, no sugarcane mapping tools are available to generate annual maps of sugarcane at the field scale over large regions. In this study, we developed a pixel- and phenology-based mapping tool to produce an annual map of sugarcane at 10-m spatial resolution by analyzing time-series Landsat-7/8, Sentinel-2 and Sentinel-1 images (LC/S2/S1) during August 31, 2017 - July 1, 2019 in Guangxi province, China, which accounts for 65% of sugarcane production of China. First, we generated annual maps of croplands and other land cover types in 2018. Second, we delineated the cropping intensity (single, double and triple cropping in a year) for all cropland pixels in 2018. Third, we identified sugarcane fields in 2018 based on its phenological characteristics. The resultant 2018 sugarcane map has producer, user and overall accuracies of 88%, 96% and 96%, respectively. According to the annual sugarcane map in 2018, there was a total of 8940 km2 sugarcane in Guangxi, which was ~1% higher than the estimate from the Guangxi Agricultural Statistics Report. Finally, we identified green-up dates of those sugarcane fields in 2019, which could be used to support the sugarcane planting and management activities. Our study demonstrates the potential of the pixel- and phenology-based sugarcane mapping tool (both the algorithms and the LC/S2/S1 time series images) in identifying croplands, cropping intensity and sugarcane fields in the complex landscapes with diverse crop types, fragmented crop fields and frequent cloudy weather. The resultant annual maps from this study could be used to assist farms and sugarcane mills for sustainable sugarcane production and environment.

Accurate estimation of gross primary production (GPP) is critical for understanding ecosystem response to climate variability and change. Satellite-based diagnostic models, which use satellite images and/or climate data as input, are widely used to estimate GPP. Many models used the Normalized Difference Vegetation Index (NDVI) to estimate the fraction of absorbed photosynthetically active radiation (PAR) by vegetation canopy (FPAR(canopy)) and GPP. Recently, the Enhanced Vegetation Index (EVI) has been increasingly used to estimate the fraction of PAR absorbed by chlorophyll (FPAR(chl)) or green leaves (FPAR(green)) and to provide more accurate estimates of GPP in such models as the Vegetation Photosynthesis Model (VPM), Temperature and Greenness (TG) model, Greenness and Radiation (GR) model, and Vegetation Index (VI) model. Although these EVI-based models perform well under non-drought conditions, their performances under severe droughts are unclear. In this study, we run the four EVI-based models at three AmeriFlux sites (rainfed soybean, irrigated maize, and grassland) during drought and non-drought years to examine their sensitivities to drought. As all the four models use EVI for FPAR estimate, our hypothesis is that their different sensitivities to drought are mainly attributed to the ways they handle light use efficiency CLUE), especially water stress. The predicted GPP from these four models had a good agreement with the GPP estimated from eddy flux tower in non-drought years with root mean squared errors (RMSEs) in the order of 2.17 (VPM), 2.47 (VI), 2.85 (GR) and 3.10 g C m(-2) day(-1) (TG). But their performances differed in drought years, the VPM model performed best, followed by the VI, GR and TG, with the RMSEs of 1.61, 232, 3.16 and 3.90 g C m(-2) day-1 respectively. TG and GR models overestimated seasonal sum of GPP by 20% to 61% in rainfed sites in drought years and also overestimated or underestimated GPP in the irrigated site. This difference in model performance under severe drought is attributed to the fact that the VPM uses satellite-based Land Surface Water Index (LSWI) to address the effect of water stress (deficit) on LUE and GPP, while the other three models do not have such a mechanism. This study suggests that it is essential for these models to consider the effect of water stress on GPP, in addition to using EVI to estimate FPAR, if these models are applied to estimate GPP under drought conditions. (C) 2015 Elsevier Inc All rights reserved.

Agricultural land use change substantially affects climate, water, ecosystems, biodiversity, and human welfare. In recent decades, due to increasing population and food demand and the backdrop of global warming, croplands have been expanding into higher latitude regions. One such hotspot is paddy rice expansion in northeast China. However, there are no maps available for documenting the spatial and temporal patterns of continuous paddy rice expansion. In this study, we developed an automated, Landsat-based paddy rice mapping (Landsat-RICE) system that uses time series Landsat images and a phenology-based algorithm based on the unique spectral characteristics of paddy rice during the flooding/transplanting phase. As a pilot study, we analyzed all the available Landsat images from 1986 to 2010 (498 scenes) in one tile (path/row 113/27) of northeast China, which tracked paddy rice expansion in epochs with five-year increments (1986-1990, 1991-1995, 1996-2000, 2001-2005, and 2006-2010). Several maps of land cover types (barren land and built-up land; evergreen, deciduous and sparse vegetation types; and water-related land cover types such as permanent water body, mixed pixels of water and vegetation, spring flooded wetlands and summer flooded land) were generated as masks. Air temperature was used to define phenology timing and crop calendar, which were then used to select Landsat images in the phenology-based algorithms for paddy rice and masks. The resultant maps of paddy rice in the five epochs were evaluated using validation samples from multiple sources, and the overall accuracies and Kappa coefficients ranged from 84 to 95% and 0.6-0.9, respectively. The paddy rice area in the study area substantially increased from 1986 to 2010, particularly after the 1990s. This study demonstrates the potential of the Landsat-RICE system and time series Landsat images for tracking agricultural land use changes at 30-m resolution in the temperate zone with single crop cultivation. (C) 2015 Elsevier Inc. All rights reserved.

article i nfo Drought affects vegetation photosynthesis and growth. Many studies have used the normalized difference vege- tation index (NDVI), which is calculated as the normalized ratio between near infrared and red spectral bands in satellite images, to evaluate the response of vegetation to drought. In this study, we examined the impacts of drought on three vegetation indices (NDVI, enhanced vegetation index, EVI, and land surface water index, LSWI) and CO2 flux from three tallgrass prairie eddy flux tower sites in the U.S. Gross primary production (GPP) was also modeled using a satellite-based Vegetation Photosynthesis Model (VPM), and the modeled GPP (GPPVPM) was compared with the GPP (GPPEC) derived from eddy covariance measurements. Precipitation at two sites in Oklahoma was 30% below the historical mean in both years of the study period (2005-2006), whilethesiteinIllinoisdidnotexperiencedroughtinthe2005-2007studyperiod.TheEVIexplainedtheseason- al dynamics of GPP better than did NDVI. The LSWI dropped below zero during severe droughts in the growing season, showing its potential to track drought. The result shows that GPP was more sensitive to drought than were vegetation indices, and EVI and LSWI were more sensitive than NDVI. We developed a modified function (Wscalar), calculated as a function of LSWI, to account for the effect of severe droughts on GPP in VPM. The GPPVPM from the modified VPM accounted for the rapid reduction in GPP during severe droughts and the seasonal dynamics of GPPVPM agreed reasonably well with GPPEC. Our analysis shows that 8-day averaged values (temper- ature,vapor-pressuredeficit)donotreflectthe short-termextreme climate events well,suggesting thatsatellite- based models may need to be run at daily or hourly scales, especially under unfavorable climatic conditions.

The changing climate is shifting the seasonality of photosynthesis in vegetation, including the start (SOS), end (EOS), and length (LOS) of the growing season, and the peak photosynthesis timing (PPT). While the SOS, EOS, and LOS have been widely investigated, the PPT of grasslands – as a proxy for the response of seasonal plant photosynthesis to climate change – has been overlooked. In this study, we propose a hybrid generalized additive model (HGAM) method to extract PPT using the Vegetation Photosynthesis Model (VPM)-based gross primary production (GPP) product, and we examine the dynamics, drivers, and consequences of PPT changes in temperate and alpine grasslands in China over 2000–2016. We found that the PPTs in temperate and alpine grasslands have exhibited advancing (with −0.68 days yr−1, p

article i nfo Due to increasing global demand for natural rubber products, rubber (Hevea brasiliensis) plantation expan- sion has occurred in many regions where it was originally considered unsuitable. However, accurate maps of rubber plantations are not available, which substantially constrain our understanding of the environmental and socioeconomic impacts of rubber plantation expansion. In this study we developed a simple algorithm for accurate mapping of rubber plantations in northern tropical regions, by combining a forest map derived from microwave data and unique phenological characteristics of rubber trees observed from multi- temporal Landsat imagery. Phenology of rubber trees and natural evergreen forests in Hainan Island, China, was evaluated using eighteen Landsat TM/ETM+ images between 2007 and 2012. Temporal profiles of the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Land Surface Water Index (LSWI), and near-infrared (NIR) reflectance for rubber trees and natural forest were constructed. The results showed that rubber plantations are distinguishable from natural evergreen forests in two phenolog- ical phases: 1) during the defoliation (leaf-off) phase in late February-March, vegetation index (NDVI, EVI, LSWI) values were lower in rubber plantations than in natural evergreen forests; and 2) during the foliation (new leaf emergence) phase in late March-April, rubber plantations had similar NDVI and LSWI values but higher EVI and NIR reflectance values than in natural forests. Therefore, it is possible to delineate rubber plan- tations within forested landscapes using one to two optical images acquired in the defoliation and/or foliation period. The mapping technique was developed and applied in the Danzhou Region of Hainan. Phased Array type L-band Synthetic Aperture Radar (PALSAR) 50-m Orthorectified Mosaic images were used to generate a forest cover map and further integrated with the phenological information of rubber plantations extracted from Landsat TM images during the foliation phase. The resultant map of rubber plantations has high accu- racy (both producer's and user's accuracy is 96%). This simple and integrated algorithm has the potential to improve mapping of rubber plantations at the regional scale. This study also shows the value of time series Landsat images and emphasizes imagery selection at appropriate phenological phase for land cover classifi

The uncertainty in tracking tropical forest extent and changes substantially affects our assessment of the consequences of forest change on the global carbon cycle, biodiversity and ecosystem services. Recently cloud-free imagery useful for tropical forest mapping from the Phased Array Type L-band Synthetic Aperture Radar (PALSAR) onboard the Advanced Land Observing Satellite (ALOS) has become available. We used PALSAR 50-m orthorectified mosaic imagery in 2009 and a decision tree method to conduct land cover classification and generate a 2009 forest map, which was evaluated using 2106 field photos from the Global Geo-referenced Field Photo Library ( http://www.eomf.ou.edu/photos ). The resulting land cover classification had a high overall accuracy of 93.3% and a Kappa Coefficient of 0.9. The PALSAR-based forest map was then compared with three existing forest cover products at three scales (regional, national, and continental): the Food and Agriculture Organization of the United Nations (FAO) Forest Resources Assessments (FRA) 2010, Global Land Cover Map with MERIS (GlobCover) 2009, and the MODIS Terra + Aqua Land Cover Type product (MCD12Q1) 2009. The intercomparison results show that these four forest datasets differ. The PALSAR-based forest area estimate is within the range (6.1–9.0 × 10 5 km 2 ) of the other three products and closest to the FAO FRA 2010 estimate. The spatial disagreements of the PALSAR-based forest, MCD12Q1 forest and GlobCover forest are evident; however, the PALSAR-based forest map provides more details (50-m spatial resolution) and high accuracy (the Producer's and the User's Accuracies were 88% and 95%, respectively) and PALSAR can be used to evaluate MCD12Q1 2009 and GlobCover 2009 forest maps. Given the higher spatial resolution, PALSAR-based forest products could further improve the modeling accuracy of carbon cycle in tropical forests.