Analiza multikolinearnosti fenoloških metrika iz modis satelitskih snimki za kukuruz i soju u Hrvatskoj

Evaluation of the potential of MODIS satellite data to predict vegetation phenology in different biomes: An investigation using ground-based NDVI measurements

Vegetation phenology is one of the key indicators for assessing inter annual changes of ecosystems while crop phenology has crucial and practical importance for precise timing of farming practices of today’s smart agriculture. The aim of this study is to develop and implement a procedure for filtering and processing daily satellite signals from Moderate Resolution Imaging Spectroradiometer (MODIS) satellites for the operational derivation of vegetation indices and land surface phenology (LSP) metrics. Second, this work investigates and selects precisely adjusted parameters for the calculation of a specific LSP metric – the SOS (start of the season). The SOS obtained from the satellite data was compared with available in situ ground measurements at 80 sites with records from 2000-2012 across the Czech Republic. The coefficients of determination, R2, for the land surface phenology based on Earth observation and traditional ground phenology indicated differences between the two methods. However, the average R2 at all sites for agricultural land was 0.31 for winter wheat onset of leaf sheath elongation; the R2 exceeded 0.5 at 21 sites and was approximately 0.7 at some sites. The developed procedure proved suitable for operational monitoring of actual crop conditions and other influence factors that are important for drought monitoring.

Trends and geographic patterns of change in vegetation phenology metrics and snowmelt timing from the MODerate resolution Imaging Spectroradiometer (MODIS) satellite data sets were analyzed for the Northern Range of Yellowstone National Park over the period 2001 to 2017. The main question posed in this analysis was “Where has the growing season length, amplitude, and integrated greenness cover changed over the past two decades on the Northern Range, particularly in relation to vegetation cover types and the timing of spring snowmelt?” Phenology metric patterns derived from the Normalized Difference Vegetation Index (NDVI) time-series at 250-m resolution were used to track changes in the growing season length, amplitude, and integrated greenness cover over the past two decades. Trend analysis showed that end of the growing season timing (EOST) and integrated greenness increased significantly over nearly 30% of the Northern Range study area, and NDVI amplitude increased significantly in several large drainages, particularly in shrub-grassland cover types. Significant variation in the start of the growing season (SOST), NDVI amplitude and integrated greenness could be further explained by the timing of spring snow melt. In years with relatively late snowmelt dates (after mid-May), higher plant growth was observed over the ensuing growing season, as captured in the amplitude and integrated greenness metrics, potentially due to elevated snow water inputs that can maintain available soil moisture levels for plant growth into the mid- and late-summer months. The ecological implications of an extended period in November–December when grassland and shrub biomass supplies remain relatively snow-free and readily accessible to grazing ungulates will require further assessments.

Forecasting tree mortality using change metrics derived from MODIS satellite data

Retrieval of Leaf Area Index in mountain grasslands in the Alps from MODIS satellite imagery

Natural hazards affect different parts of living organisms. One of these hazards is aerosols. Addressing this issue in areas that suffer from this hazard is of critical importance. Scientific research over the past two decades has shown that aerosol particles are one of the main pollutants from the perspective of public health and health. The purpose of the present study is to monitor and model aerosol temporal and spatial variations in Iran. Aerosols or airborne particles with health effects such as heart, vascular and respiratory diseases are associated. For this purpose, MODIS (or Moderate Resolution Imaging Spectroradiometer) and NOAA (or National Oceanic and Atmospheric Administration) satellite data and BTD (or Brightness Temperature Difference) were used. Given the 20-year study period (2000-2019), the output of much satellite data was divided into four five-year periods to monitor aerosols with high accuracy. The results showed that aerosol values in terms of intensity and frequency of optical depth (AOD) increased over the 20-year time series, and its intensity was higher in the last 5years. According to the results obtained from the NODIS and NOAA satellite data and comparing their outputs, NOAA satellite data were associated with outliers, which was significantly different from the other cohort data. For this reason, the MODIS satellite image output was used to monitor aerosol images. The innovation of the present study is the use of remote sensing science to monitor the effects of aerosols on the environment and human's health. According to the results from MODIS satellite data, the maximum optical depth (AOD) of the aerosols is for July 2003 with a value of 0.63, but according to the NOAA satellite data output, the maximum optical depth (AOD) of the aerosols is for March 2013 with a value of 2.54. As the values of aerosols increase in frequency and intensity and the areas with the highest intensity of aerosols have been identified, it can be overcome by careful planning of their problems. Areas, where the amount and volume of aerosols were higher, should be observed in health protocols. By preventing and maintaining good hygiene, the negative effects of aerosols can be reduced.

Vegetation indices have been commonly used over the past 30 years for studying vegetation characteristics using images collected by remote sensing satellites. One of the most commonly used is the Normalized Difference Vegetation Index (NDVI). The various stages that green vegetation undergoes during a complete growing season can be summarized through time-series analysis of NDVI data. The analysis of such time-series allow for extracting key phenological variables or metrics of a particular season. These characteristics may not necessarily correspond directly to conventional, ground-based phenological events, but do provide indications of ecosystem dynamics. A complete list of the phenological metrics that can be extracted from smoothed, time-series NDVI data is available in the USGS online resources (http://phenology.cr.usgs.gov/methods_deriving.php).This work aims to develop an open source application to automatically extract these phenological metrics from a set of satellite input data. The main advantage of QGIS for this specific application relies on the easiness and quickness in developing new plug-ins, using Python language, based on the experience of the research group in other related works. QGIS has its own application programming interface (API) with functionalities and programs to develop new features. The toolbar developed for this application was implemented using the plug-in NDVIToolbar.py. The user introduces the raster files as input and obtains a plot and a report with the metrics. The report includes the following eight metrics: SOST (Start Of Season – Time) corresponding to the day of the year identified as having a consistent upward trend in the NDVI time series; SOSN (Start Of Season – NDVI) corresponding to the NDVI value associated with SOST; EOST (End of Season – Time) which corresponds to the day of year identified at the end of a consistent downward trend in the NDVI time series; EOSN (End of Season – NDVI) corresponding to the NDVI value associated with EOST; MAXN (Maximum NDVI) which corresponds to the maximum NDVI value; MAXT (Time of Maximum) which is the day associated with MAXN; DUR (Duration) defined as the number of days between SOST and EOST; and AMP (Amplitude) which is the difference between MAXN and SOSN. This application provides all these metrics in a single step. Initially, the data points are interpolated using a moving average graphic with five and three points. The eight metrics previously described are then obtained from the spline using numpy functions. In the present work, the developed toolbar was applied to MODerate resolution Imaging Spectroradiometer (MODIS) data covering a particular region of Portugal, which can be generally applied to other satellite data and study area. The code is open and can be modified according to the user requirements. Other advantage in publishing the plug-ins and the application code is the possibility of other users to improve this application.

Phenology is an indicator of crop growth conditions, and is correlated with crop yields. In this study, a phenological approach based on a remote sensing vegetation index was explored to predict the yield in 314 counties within the US Corn Belt, divided into semi-arid and non-semi-arid regions. The Moderate Resolution Imaging Spectroradiometer (MODIS) data product MOD09Q1 was used to calculate the normalized difference vegetation index (NDVI) time series. According to the NDVI time series, we divided the corn growing season into four growth phases, calculated phenological information metrics (duration and rate) for each growth phase, and obtained the maximum correlation NDVI (Max-R2). Duration and rate represent crop growth days and rate, respectively. Max-R2 is the NDVI value with the most significant correlation with corn yield in the NDVI time series. We built three groups of yield regression models, including univariate models using phenological metrics and Max-R2, and multivariate models using phenological metrics, and multivariate models using phenological metrics combined with Max-R2 in the whole, semi-arid, and non-semi-arid regions, respectively, and compared the performance of these models. The results show that most phenological metrics had a statistically significant (p < 0.05) relationship with corn yield (maximum R2 = 0.44). Models established with phenological metrics realized yield prediction before harvest in the three regions with R2 = 0.64, 0.67, and 0.72. Compared with the univariate Max-R2 models, the accuracy of models built with Max-R2 and phenology metrics improved. Thus, the phenology metrics obtained from MODIS-NDVI accurately reflect the corn characteristics and can be used for large-scale yield prediction. Overall, this study showed that phenology metrics derived from remote sensing vegetation indexes could be used as crop yield prediction variables and provide a reference for data organization and yield prediction with physical crop significance.

Mapping abandoned agriculture with multi-temporal MODIS satellite data

Reference evapotranspiration (ETo) is an important variable in hydrological modeling, which is not always available, especially for ungauged catchments. Satellite data, such as those available from the MODerate Resolution Imaging Spectroradiometer (MODIS), and global datasets via the European Centre for Medium Range Weather Forecasts (ECMWF) reanalysis (ERA) interim and National Centers for Environmental Prediction (NCEP) reanalysis are important sources of information for ETo. This study explored the seasonal performances of MODIS (MOD16) and Weather Research and Forecasting (WRF) model downscaled global reanalysis datasets, such as ERA interim and NCEP-derived ETo, against ground-based datasets. Overall, on the basis of the statistical metrics computed, ETo derived from ERA interim and MODIS were more accurate in comparison to the estimates from NCEP for all the seasons. The pooled datasets also revealed a similar performance to the seasonal assessment with higher agreement for the ERA interim (r = 0.96, RMSE = 2.76 mm/8 days; bias = 0.24 mm/8 days), followed by MODIS (r = 0.95, RMSE = 7.66 mm/8 days; bias = −7.17 mm/8 days) and NCEP (r = 0.76, RMSE = 11.81 mm/8 days; bias = −10.20 mm/8 days). The only limitation with downscaling ERA interim reanalysis datasets using WRF is that it is time-consuming in contrast to the readily available MODIS operational product for use in mesoscale studies and practical applications.

This study is devoted to assessing the accuracy of determining the Caspian Sea surface temperature (SST) using MODIS (Moderate Resolution Imaging Spectroradiometer) satellite data. The authors used in situ temperature measurements obtained from Surface Velocity Program Barometer (SVPB) drifters deployed in the Caspian Sea in 2006 – 2007 and 2008 as part of the MACE Project. For comparison, daily MODIS LST products with a spatial resolution of 1 km, available on the Google Earth Engine platform, were used. Products obtained using the Split-Window (MOD11A1 and MYD11A1) and Temperature Emissivity Separation (TES) algorithms (MOD21A1D, MYD21A1D, MOD21A1N, MYD21A1N) for day and night surveys were analyzed. The results showed a significant linear relationship between the MODIS satellite data and drifter measurements for all considered products. For the products based on the Split-Window Algorithm, the correlation coefficients were 0.94 – 0.96, and the determination coefficients were 0.73 – 0.78. The average error for daytime data was -1.23 – -1.52 °C, and for nighttime data -2.14 – -2.16 °C. It was noted that for daytime data of the MOD11A1 and MYD11A1 products, the deviation increases with increasing water temperature. The products based on the TES algorithm showed higher accuracy. For daytime products MOD21A1D and MYD21A1D, the correlation coefficient was 0.91 – 0.93, the determination coefficient was 0.83 – 0.84, and the average error was 0.37 – 0.64 °C. For the nighttime MOD21A1N and MYD21A1N products, the correlation coefficient was 0.96, the determination coefficient was 0.91 – 0.92, and the average error was 0.03 – 0.25 °C. The MODIS data were also compared with Landsat Level-2 data. The determination coefficient was 0.92 in both cases (using MODIS Split-Window and TES products). On average, the Landsat temperature was higher by 0.5 °C when compared with the MYD11A1 and MOD11A1 products (daytime) and lower by 0.85 °C when compared with the MYD21A1D and MOD21A1D products. Overall, the study showed that MODIS products demonstrate high accuracy in reproducing the Caspian Sea surface temperature, with products using the TES algorithm having higher accuracy compared to products based on the Split-Window algorithm.

Trends and geographic patterns of change in vegetation phenology metrics and snowmelt timing from the MODerate resolution Imaging Spectroradiometer (MODIS) satellite data sets were analyzed across the state of Alaska for all wildfires that burned during the years 2004 and 2005. Phenology metric patterns (over the period 2000 to 2018) derived from the normalized difference vegetation index (NDVI) time-series at 250-m resolution tracked changes in the growing season length and integrated greenness cover over the past two decades. NDVI metrics showed that end of the growing season timing (EOST) and integrated greenness increased significantly in the majority of severely burned areas in Interior Alaska over the past decade particularly in low elevation zones below 500 m. In years with relatively early snowmelt dates (3 to 10 days earlier than the long-term mean), lower plant growth was observed over the ensuing growing season, potentially due to lower snow water inputs that can maintain available soil moisture levels for plant growth into the mid- and late-summer months. Statewide trends in MODIS phenology metrics indicate that the predominant vegetation cover type in most regrowing burned areas of Interior Alaska a decade post-fire is still deciduous shrub and young tree cover. Several large fires of special interest were identified to monitor with remote sensing and continue to track long-term recovery patterns and rates.

Retrieval of the seawater reflectance for suspended solids monitoring in the East China Sea using MODIS, MERIS and GOCI satellite data

Understanding trends in vegetation phenology and growing season productivity at a regional scale is important for global change studies, particularly as linkages can be made between climate shifts and the vegetation’s potential to sequester or release carbon into the atmosphere. Trends and geographic patterns of change in vegetation growth and phenology from the MODerate resolution Imaging Spectroradiometer (MODIS) satellite data sets were analyzed for the state of Alaska over the period 2000 to 2018. Phenology metrics derived from the MODIS Normalized Difference Vegetation Index (NDVI) time-series at 250 m resolution tracked changes in the total integrated greenness cover (TIN), maximum annual NDVI (MAXN), and start of the season timing (SOST) date over the past two decades. SOST trends showed significantly earlier seasonal vegetation greening (at more than one day per year) across the northeastern Brooks Range Mountains, on the Yukon-Kuskokwim coastal plain, and in the southern coastal areas of Alaska. TIN and MAXN have increased significantly across the western Arctic Coastal Plain and within the perimeters of most large wildfires of the Interior boreal region that burned since the year 2000, whereas TIN and MAXN have decreased notably in watersheds of Bristol Bay and in the Cook Inlet lowlands of southwestern Alaska, in the same regions where earlier-trending SOST was also detected. Mapping results from this MODIS time-series analysis have identified a new database of localized study locations across Alaska where vegetation phenology has recently shifted notably, and where land cover types and ecosystem processes could be changing rapidly.

Vegetation Health Index is computed based on the vegetation index and land surface temperature which are very important satellite products for drought monitoring with remote sensing. The quality of Temperature Condition Index and Vegetation Health Index deeply depends on the quality of the Land Surface Temperature (LST). The present retrieval algorithm of Moderate Resolution Imaging Spectroradiometer (MODIS) LST products is quite good globally but for a local area there is still a room to improve by taking the effects of aerosol, water vapor, view zenith and observing time into account. The paper compared the MODIS LST with the observed surface temperature in weather station in the farming area of Eastern China. The correlations between the MODIS LST and the MODIS TPW, MODIS AOD, observing time, observing angle and DEM, respectively, were analyzed. And finally a regression equation to correct the temperature difference is formed and used to correct the MODIS LST retrieved from MODIS satellite. The results showed that the corrected LST is close to the observed with the decreased root mean square error and the increased correlation coefficient. The Vegetation Health Index was further calculated and used to monitoring the winter wheat drought in Eastern China. The drought situation was better captured by the improved VHI.