During pre-processing, this DN can be transformed, calibrated, and filtered, to understand different biophysical parameters of that space. There are a number of different constraints that must be accounted for when determining soil water content . Depending on how the DL is transformed, we accounted for those constraints. The DN given by sentinel one data can be transformed into alpha, beta, orgamma values. The team explored these different transformations to determine what was best for determining AWD adoption in the MRD. Because of the complexity of surface roughness and the rice crop itself, the team decided on a beta transformation, using beta values to create a wetness index that represents soil-water content . First, the β° transformation takes the DN and transforms it into reflectivity data. We then calibrate β° using a script embedded in the downloadable file via the ESA web interface for the beta calibration . Calibration for beta values mostly uses a constant unique to the acquisition date, time, and incidence angle, so the shift in distribution after calibration was not as significant as it might be for alpha. After calibration, the data was run through a median smoothing filter to remove speckling . These pre-processing steps created a multi-temporal dataset ready for analysis.The test plots in An Giang, Dong Thap, and Ben Tre were used to train our model. We calculated the maximum and minimum values for the VV and VH bands for each soil type, greenhouse tables with the minimum representing a “flooded” field with low reflectivity and maximum representing a “dry” field with high reflectivity . Because the maximum and minimum VV and VH values were not significantly different between soil types, we used the maximum and minimum from An Giang province moisture meters.

Using these values, we excluded VV and VH values below the minimum and above the maximum of those rice paddies across the entire dataset. This process, in essence, acted as a high pass filter to pull out water features, urban features, and other nonpaddy features .SAR data shows promise as a tool to document adoption of water-saving practices. Results of this study support a further research and development to use remotely sensed data for documenting adoption of AWD. Specifically for alluvial soil types, as shown with the parallel patterns between the Wetness Index and in situ moisture meters, this approach shows promising links to real world conditions. The satellite data used, Sentinel-1, had a moderately useful return time for analysis of AWD adoption in the MRD. Each satellite returns at 12-day intervals, meaning that the rate of return for the two satellites combined is every six days. While strict adherence to AWD technology would dictate that adoption means dry down of rice fields every five days, a pattern of dry down was still detectable every sixe to 12 days. Therefore, we believe that farmers are not necessarily utilizing a five day dry down period, but perhaps a longer period of time between irrigations. The SAR data analysis is much more cost effective than other efforts to measure the extent of AWD adoption. This research project conducted 285 household surveys in four delta provinces for approximately the same cost of developing and executing an analysis across the entire rice growing Mekong River Delta. With further development, this cost would continue to decline to near zero, using free data.There are several issues with the data availability and approach discussed in this paper. Estimating AWD adoption using remotely sensed data is a complex andnuanced issue. Not only was data availability an issue, but also the complexity of SAR data itself. We suggest two avenues for correcting data availability and data analysis issues. Further analysis of SAR data with additional time series datasets could yield more accurate results.

The GIS team originally discussed using ASTER data. The rate of return for each satellite pass-over was determined to be insufficient to detect AWD at use. Therefore, this data was not used in analysis of AWD adoption in the MRD. To understand the sinusoidal patterns of each cell in the raster data sets, we suggest a power spectral density test. The average dry down, while a good estimate of an overall likelihood of some degree of AWD adoption, does not necessarily account for the complex periodicity of the SAR data across the delta over time. The change detection approach used for the SAR data analysis illustrates the average change over time for each cell in the tile that covers the MRD. However, a more nuanced approach would be to use a power spectral density analysis to understand the periodicity of the “dry down” signal and the power of that signal. In other words, cells with increased likelihood of exhibiting an oscillating habit , would score higher on the power spectral density analysis. Second, we suggest that researchers repeat the same analysis for all three seasons of rice to increase the accuracy of the power spectral density test. While the approach was able to detect patterns of “dry down” across the delta, conclusions would create a more robust signal analysis with longer sets of continuous data. If each growing season in a year could be analyzed continuously, especially over a series of years, this would enable detection of a more conclusive and reliable AWD signal.The story of Vietnam’s post-war rural landscape is still unfolding. Vietnam spent 25 years mired in conflict to free themselves of colonial ties, only to start a new battle: the “irrigation front” . While American-led nation building efforts in the 1950s included a heavy focus on technology transfer and adoption , it was in the post-war period after reunification in which Vietnam built the bulk of modern irrigation and infrastructure, shifting the landscape from a traditional and flexible adaptive environment to a technologically robust and controlled deltaic system . Beginning in the 1980s, a cascade of physical and social changes rippled across the Vietnamese Mekong River Delta . The canal and sluice gate system tamed the Mekong and the surrounding delta and, combined with adoption of Green Revolution philosophies of input dependent production, allowed Vietnam to emerge as a rice production giant .

Between 1990 and 2010 Vietnam became one of the region’s leading rice producers. The gross domestic product per capita rose from $98 in 1990 to $2,052 in 2014 . Simultaneously, structural reformations in the late 1980s, known as Doi Moi, reorganized how land was distributed to individuals and privilege male ownership and inheritance .Farmers felt the impact of these larger environmental, economic, and social changes at the household level. The majority of farms in Vietnam consist of two or three small plots of land, totaling less than one hectare . Over time, these small plots shifted to solely rice production, instead of vegetables or livestock outside the wet season . Further, rice production is typically a man’s job, leaving women with decreasing options for local agricultural labor jobs and pushing them to migrate to the city . Most farmers in the Mekong River Delta of Vietnam must keep pace with an increasing demand for rice, which now requires three rice crops per year, or what local experts call the triple rice crop. This primarily male burden will be referred to in this chapter as the “triple crop burden”1 – the socioeconomic pressure to produce three rice crops annually. Triple rice cropping is primarily accomplished using improved seeds, fertilizers, and pesticides, which we refer to as conventional intensification in this chapter. In the late 2000s, the Ministry of Agriculture and Rural Development began pushing a “sustainable intensification” approach, to keep steady the triple ricecrop while making production practices more environmentally friendly . Research on the gendered implications of this policy is quite new, and consists primarily of white papers and project reports. There are three gaps in the literature on how gender influences farming practice adoption in Vietnam. First, because of Vietnam’s recent success in rice production using CI methods, and the subsequent policies to reduce environmental impacts through SI practices, there are few studies outlining the current suite of CI and SI practices used by MRD farmers. Second, flood tray despite the net increase in rice farming over all other crops in the delta, there is a dearth of studies examining how variance in household livelihoods impacts gendered CI and SI practice adoption in Vietnam. Third, there is a lack of understanding in how the country’s increasingly female migrant population, who sends remittances back to the farm, influences adoption of CI versus SI methods. The purpose of this research is to evaluate these three facets of the SI movement in Vietnam. We conducted gender disaggregated plot-level study in Tien Giang Province, in the heart of the triple cropping delta, using household surveys to explore gendered differences in on-farm practices. We evaluated how farmers differentially adopt SI and CI methods, how household characteristics influence male and female adoption of these practices, and to what extent remittances influence men and women’s plots. This research adopts an agroecology framework to understand gendered livelihood strategies on the farm. The discipline of agroecology analyzes the sustainability and robustness of food systems. Through altering management practices and on-farm goals, agroecology is a distinct form of agricultural research both philosophically and materially. First, agroecology includes “facilitating” or “managing” processes rather than “controlling” them . Second, it treats components of the system as interrelated rather than separate, sometimes called the “environmental complex” . Third, it accounts for complex human interactions with the food system in addition to economic viability of the farm, such as equality and fair labor practices for on-farm labor and long-term environmental impacts of production processes .

To gather locally based data that links to larger geopolitical and agricultural processes, agroecology often uses the livelihoods framework, which aims to find institutional and organizational avenues to decrease poverty through access to resources and strategies or capabilities . In agroecology, livelihood research aims to understand how populations dependent on natural resources for their livelihoods are subject to wealth inequality and that diversifying asset bundles to absorb shocks, including diversifying cropping systems or income streams, can improve household stability . In this research, we treat migration and access to markets as capabilities in this livelihood bundle, and explore the household capitals to understand gendered practice adoption.The chapter proceeds in the following fashion. The first part of the chapter outlines the history of intensification of rice production in the Mekong River Delta of Vietnam, as well as the more recent efforts to downgrade use of inputs while sustaining yields. We also discuss how decisions to adopt more or less conventional intensification practices can be a gendered phenomenon. The second part of the chapter states the methods for data collection and analysis in this study. The third part outlines results of the household surveys and multivariate probit analysis. The fourth section discusses the implications of this research for sustainable intensification in Vietnam, gendered resource control, and female inclusion in trainings and decision making. We conclude with several policy recommendations and future directions for research.The Green Revolution was an international effort beginning in the 1960s to increase food security through agricultural intensification. Intensification is the process of increasing productivity from the same unit area of land. The Green Revolution achieved intensification through three primary efforts: improved seed varieties, mechanization of labor, and chemical inputs . In Vietnam, the Green Revolution did not take hold as early as other countries due to the war . In the Mekong River Delta of Vietnam, Green Revolution practices were ubiquitously adopted, including the extensive irrigation network and improved seed varieties. Irrigation and genetically modified seeds slowly became the standard in what we conceive of as conventional intensification agriculture in this chapter, which allows three rice crops per year. The plough and combine harvester are the primary machinery used in crop production in the MRD. Chemical inputs, such as pesticides and fertilizers, become necessary in a system that is highly productive for three crops per year due to nutrient depletion and increase pest pressure. The Green Revolution was touted as a solution to global food security issues, but criticized by many for its environmental and social costs. It has been criticized for being shortsighted and environmentally risky . With rapid urbanization and population growth, there was increased concern in the 1980s of expansion of agricultural land competing with conservation and food security efforts . In other words, conversion of native ecosystems to agricultural land was ecologically unsustainable.