Highlights from the Literature: January to March 2017

Apr 26, 2017

Data products distributed by the Land Processes (LP) Distributed Active Archive Center (DAAC) are used in many different applications. They play an important role in modeling, help to detect changes to the landscape, and are a way to assess ecosystem variables, to name a few. Three of those applications, published between January and March 2017, are highlighted below. A more exhaustive, albeit not complete, list can be accessed via the LP DAAC Publications webpage.

Sheng, Y., Liu, X., Yang, X., Xin, Q., Deng, C., and Li, X., 2017, Quantifying the spatial and temporal relationship between air and land surface temperatures of different land-cover types in Southeastern China: International Journal of Remote Sensing, v. 38, no. 4, p. 1114–1136. [Also available at http://dx.doi.org/10.1080/01431161.2017.1280629].
In this paper, Sheng and others (2017) explore the relationship between near-surface air temperature, observed through automatic weather stations 2 meters (m) above ground, and daytime and nighttime Terra Moderate Resolution Imaging Spectroradiometer (MODIS) daily Land Surface Temperature (LST) (MOD11A1) Version 5 data. The authors use statistical analysis to study the difference between the two measurements for different land cover types in the northern part of the Zhejiang Province of Southeast China from January 1, 2014, to December 31, 2014. The land cover types are determined by weather station locations and then are classified as impervious surface, water, or vegetation. The authors use the MODIS Reprojection Tool (MRT) to re-project and mosaic the Terra MODIS LST data. Several tests are performed to determine if there is a relationship between air temperature and LST, to observe the effect of seasonality, to understand the correlation with spatial distribution and altitude (below 1,200 m) of the weather stations, and to discover the optimal spatial extent where LST agrees with air temperature. The authors find there is a better agreement of LST and air temperature in nighttime measurements. Relative to season, they find vegetated areas show good agreement between LST and air temperature and that the hot season, May to September, has the weakest correlation. Overall, the correlation coefficient between LST and air temperature varies greatly between different seasons and different land types, but seems to have no relationship with altitude. The authors state that results from this study will aid in building a model to derive air temperature from remotely sensed data in the future. They also believe future research could improve the results by observing MODIS data at different overpassing times, obtaining air surface temperature data from different years, and studying higher elevation areas.

Brown, M., and McCarty, J., 2017, Is remote sensing useful for finding and monitoring urban farms?: Applied Geography, v. 80, p. 23–33. [Also available at http://dx.doi.org/10.1016/j.apgeog.2017.01.008].
In this study, Brown and McCarty (2017) use 16-day Terra MODIS Vegetation Indices Normalized Difference Vegetation Index (NDVI) (MOD13Q1) data, along with Landsat 8 data, to study if remote sensing data can find and monitor urban farms. The authors also incorporate data mined from social media and online farming networks. The study focuses on the cities of Detroit, Michigan, USA; Ho Chi Minh City, Vietnam; Harare, Zimbabwe; and Dakar, Senegal, for the period January 1, 2012, to January 1, 2016. Terra MODIS NDVI data were used primarily to increase the temporal resolution from every 32 days with Landsat 8 composites to 16 days. First, the authors mine data to identify urban farms and greenspaces in the four cities and then delineate the city boundaries in Google Earth Pro to determine areas of residential structures and roads near the urban farms. The authors compare patterns, textures, and colors of known urban farms to parks and semi-arid landscapes to differentiate between proposed urban farms and nearby greenspaces. Next, the authors create a multi-year time series for each urban farm and greenspace polygon using the Landsat 8 and MODIS data to calculate a mean value for each polygon. All four study areas show some difference in mean vegetation data when comparing urban farms to non-urban farm greenspaces. The authors find the Terra MODIS satellite sensor performs better than the Landsat 8 satellite and identifies more urban farms over non-urban farm greenspace in Ho Chi Minh City and also performs better in monitoring known urban farms in the city of Harare. The authors find Terra MODIS and Landsat 8 data are capable of monitoring known urban farms, but NDVI data are not ideal for discovering urban farms within the four cities. With this mixed method approach the authors declare that they are able to gain much more information through the remote sensing data than otherwise would be possible from strictly in-situ data. However, they also state that more in-situ data are needed to aid satellite sensors in monitoring additional urban farms. For future studies, the authors believe remote sensing can be a valuable tool for the urban farm community to identify, monitor, and determine where future urban farms should be placed to be the most beneficial to local communities.

Mohammadi, A., Costelloe J. F., and Ryu, D., 2017, Application of time series of remotely sensed normalized difference water, vegetation and moisture indices in characterizing flood dynamics of large-scale arid zone floodplains: Remote Sensing of Environment, v. 190, p. 70–82. [Also available at http://dx.doi.org/10.1016/j.rse.2016.12.003].
In this study, Mohammadi and others (2017) use Terra MODIS daily surface reflectance Version 6 data (MOD09GA) and Landsat 5 Thematic Mapper (TM) data to demonstrate how spatiotemporal floodplains respond to floods in large arid regions. Here, the authors study the large arid river system in the Cooper Creek floodplain of Queensland and South Australia from 2000 to 2012. Pixel-based time-series of water, vegetation, and moisture indices are created from various bands of the Terra MODIS surface reflectance data. The authors specifically use MODIS data to provide a detailed daily time-series of flood images for a single flood event. The authors note that the Terra MODIS data complements the Landsat 5 TM data by providing more data, and thus more opportunity for cloud-free acquisitions. The authors use the MODIS Reprojection Tool (MRT) to mosaic, re-project, and clip the data to the Cooper Creek floodplain. The Normalized Difference Vegetation Index (NDVI) is calculated using MODIS Bands 1 and 2 to observe changes in vegetation, and Bands 4 and 6 are used to calculate a modified Normalized Difference Water Index (mNDWI) to discern water from vegetation and soil. Lastly, a Land Surface Water Index (LSWI) is calculated using Bands 2 and 6 to differentiate between dry and wet regions. The authors note that studying vegetation and water at the same time allows for measurement of lag time between flooding and peak vegetation, as well as how long surface water and green vegetation persist—two factors that are key for studying arid zone floodplain behavior. Using the Terra MODIS data, the authors find the large extent and high frequency of the data provide advantages for characterizing the changes in inundation for the large-scale floodplains during times of instant daily flooding, as opposed to total cumulative flood extent. They are able to observe that after a flooding pulse the mNDWI and LSWI values peak, while the NDVI values drop to a minimum, and the average lag time between peak flooding and peak vegetation productivity is 41 days. The authors believe that while data from the Sentinel-2, Landsat, and SPOT satellites, along with data from commercial satellites, will be able to provide high spatial and temporal resolution in the future, the consistent, high-frequency of data from Terra MODIS will still be able to provide a highly relevant and essential historical dataset for large floodplains. The authors state that this study’s approach is globally suitable to study other large, arid zone, low-gradient floodplains.

Material written by Danielle Golon​¹

¹ Innovate!, Inc., contractor to the U.S. Geological Survey, Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota, USA. Work performed under USGS contract G15PD00403 for LP DAAC².

² LP DAAC Work performed under NASA contract NNG14HH33I.