Phenology Detection Research Articles

Analysis of Differences in Phenology Extracted from the Enhanced Vegetation Index and the Leaf Area Index.

Remote-sensing phenology detection can compensate for deficiencies in field observations and has the advantage of capturing the continuous expression of phenology on a large scale. However, there is some variability in the results of remote-sensing phenology detection derived from different vegetation parameters in satellite time-series data. Since the enhanced vegetation index (EVI) and the leaf area index (LAI) are the most widely used vegetation parameters for remote-sensing phenology extraction, this paper aims to assess the differences in phenological information extracted from EVI and LAI time series and to explore whether either index performs well for all vegetation types on a large scale. To this end, a GLASS (Global Land Surface Satellite Product)-LAI-based phenology product (GLP) was generated using the same algorithm as the MODIS (Moderate Resolution Imaging Spectroradiometer)-EVI phenology product (MLCD) over China from 2001 to 2012. The two phenology products were compared in China for different vegetation types and evaluated using ground observations. The results show that the ratio of missing data is 8.3% for the GLP, which is less than the 22.8% for the MLCD. The differences between the GLP and the MLCD become stronger as the latitude decreases, which also vary among different vegetation types. The start of the growing season (SOS) of the GLP is earlier than that of the MLCD in most vegetation types, and the end of the growing season (EOS) of the GLP is generally later than that of the MLCD. Based on ground observations, it can be suggested that the GLP performs better than the MLCD in evergreen needleleaved forests and croplands, while the MLCD performs better than the GLP in shrublands and grasslands.

Estimation of SOS and EOS for Midwestern US Corn and Soybean Crops

Understanding crop phenology is fundamental to agricultural production, management, planning, and decision-making. This study used 250 m 16-day Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI) time-series data to detect crop phenology across the Midwestern United States, 2007–2015. Key crop phenology metrics, start of season (SOS) and end of season (EOS), were estimated for corn and soybean. For such a large study region, we found that MODIS-estimated SOS and EOS values were highly dependent on the nature of input time-series data, analytical methods, and threshold values chosen for crop phenology detection. With the entire sequence of MODIS EVI time-series data as input, SOS values were inconsistent compared to crop emergent dates from the United States Department of Agriculture (USDA) Crop Progress Reports (CPR). However, when we removed winter EVI images from the time-series data to reduce impacts of snow cover, we obtained much more consistent SOS estimation. Various threshold values (10 to 50% of seasonal EVI amplitude) were applied to derive SOS values. For both corn’s and soybean’s SOS estimation, a threshold value of 25% generated the best overall agreement with the CPR crop emergent dates. Root-mean-square error (RMSE) values were 4.81 and 5.30 days for corn and soybean, respectively. For corn’s EOS estimation, a threshold value of 40% led to a high R2 value of 0.82 and RMSE value of 5.16 days. We further examined spatial patterns of SOS and EOS for both crops—SOS for corn displayed a clear south-north gradient; the southern portion of the Midwest US has earlier SOS and EOS dates.

Using data from Landsat, MODIS, VIIRS and PhenoCams to monitor the phenology of California oak/grass savanna and open grassland across spatial scales

The Mediterranean-type oak/grass savanna of California is composed of widely spaced oak trees with understory grasses. These savanna regions are interspersed with large areas of more open grasslands. The ability of remotely sensed data (with various spatial resolutions) to monitor the phenology in these water-limited oak/grass savannas and open grasslands is explored over the 2012–2015 timeframe using data from Landsat (30m), the MODerate resolution Imaging Spectroradiometer (MODIS – gridded 500m), and the Visible Infrared Imaging Radiometer Suite (VIIRS – gridded 500m) data. Vegetation phenology detected from near-ground level, webcam based PhenoCam imagery from two sites in the Ameriflux Network (long-term flux measurement network of the Americas) (Tonzi Ranch and Vaira Ranch) is upscaled, using a National Agriculture Imagery Program (NAIP) aerial image (1m), to evaluate the detection of vegetation phenology of these savannas and grasslands with the satellite data. Results show that the Normalized Difference Vegetation Index (NDVI) time series observed from the satellite sensors are all strongly correlated with the PhenoCam NDVI values from Tonzi Ranch (R2>0.67) and Vaira Ranch (R2>0.81). However, the different viewing geometries and spatial coverage of the PhenoCams and the various satellite sensors may cause differences in the absolute phenological transition dates. Analysis of frequency histograms of phenological dates illustrate that the phenological dates in the relatively homogeneous open grasslands are consistent across the different spatial resolutions, in contrast, the relatively heterogeneous oak/grass savannas display has somewhat later greenup, maturity, and dormancy dates at 30m resolution than at 500m scale due to the different phenological cycles exhibited by the overstory trees and the understory grasses. In addition, phenologies derived from the MODIS view angle corrected reflectance (Nadir BRDF-Adjusted Reflectance – NBAR) and the newly developed VIIRS NBAR are shown to provide comparable phenological dates (majority absolute bias ≤2days) in this area.

Detection of different vegetation phenology of artificial Japanese larch forest and deciduous broad-leaved forest in Hokkaido eastern region by Terra/MODIS in the defoliation period

Normalized Difference Water Index (NDWI), Green-Red Vegetation Index (GRVI), and other indices of remotely detected spectrum combinations have been proposed in order to evaluate the vegetation changes and distributions. The training area of artificial Japanese larch forest and deciduous broad leaved forest (Japanese white birch and Japanese oak) in Hokkaido were extracted from MODIS data based on RapidEye data and GIS data of Vegetation Survey on 6/7th National Basic Survey on Natural Environment (Biodiversity Center of Japan, Ministry of the Environment (http://www.biodic.go.jp/index_e.html)), the mean elevation, and the monthly mean temperature. The NDWI and GRVI were obtained for each training area from MODIS data. As a result, i) the MODIS GRVI of artificial Japanese larch forest were around -0.1, while the MODIS NDWI were greater than 0.5, ii) the MODIS GRVI of deciduous broad leaved forest were around 0.0, while the MODIS NDWI were less than 0.5, iii) these MODIS GRVI of Japanese larch and deciduous broad leaved forest were similar to GRVI of yellow and red leaves of larch and leafless birch calculated from the camera image of the canopy.

Crop Phenology Detection Using High Spatio-Temporal Resolution Data Fused from SPOT5 and MODIS Products.

Timely and efficient monitoring of crop phenology at a high spatial resolution are crucial for the precise and effective management of agriculture. Recently, satellite-derived vegetation indices (VIs), such as the Normalized Difference Vegetation Index (NDVI), have been widely used for the phenology detection of terrestrial ecosystems. In this paper, a framework is proposed to detect crop phenology using high spatio-temporal resolution data fused from Systeme Probatoire d'Observation de la Tarre5 (SPOT5) and Moderate Resolution Imaging Spectroradiometer (MODIS) images. The framework consists of a data fusion method to produce a synthetic NDVI dataset at SPOT5’s spatial resolution and at MODIS’s temporal resolution and a phenology extraction algorithm based on NDVI time-series analysis. The feasibility of our phenology detection approach was evaluated at the county scale in Shandong Province, China. The results show that (1) the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) algorithm can accurately blend SPOT5 and MODIS NDVI, with an R2 of greater than 0.69 and an root mean square error (RMSE) of less than 0.11 between the predicted and referenced data; and that (2) the estimated phenology parameters, such as the start and end of season (SOS and EOS), were closely correlated with the field-observed data with an R2 of the SOS ranging from 0.68 to 0.86 and with an R2 of the EOS ranging from 0.72 to 0.79. Our research provides a reliable approach for crop phenology mapping in areas with high fragmented farmland, which is meaningful for the implementation of precision agriculture.

Monitoring of winter wheat distribution and phenological phases based on MODIS time-series: A case study in the Yellow River Delta, China

Accurate winter wheat identification and phenology extraction are essential for field management and agricultural policy making. Here, we present mechanisms of winter wheat discrimination and phenological detection in the Yellow River Delta (YRD) region using moderate resolution imaging spectroradiometer (MODIS) time-series data. The normalized difference vegetation index (NDVI) was obtained by calculating the surface reflectance in red and infrared. We used the Savitzky-Golay filter to smooth time series NDVI curves. We adopted a two-step classification to identify winter wheat. The first step was designed to mask out non-vegetation classes, and the second step aimed to identify winter wheat from other vegetation based on its phenological features. We used the double Gaussian model and the maximum curvature method to extract phenology. Due to the characteristics of the time-series profiles for winter wheat, a double Gaussian function method was selected to fit the temporal profile. A maximum curvature method was performed to extract phenological phases. Phenological phases such as the green-up, heading and harvesting phases were detected when the NDVI curvature exhibited local maximum values. The extracted phenological dates then were validated with records of the ground observations. The spatial patterns of phenological phases were investigated. This study concluded that, for winter wheat, the accuracy of classification is 87.07%, and the accuracy of planting acreage is 90.09%. The phenological result was comparable to the ground observation at the municipal level. The average green-up date for the whole region occurred on March 5, the average heading date occurred on May 9, and the average harvesting date occurred on June 5. The spatial distribution of the phenology for winter wheat showed a significant gradual delay from the southwest to the northeast. This study demonstrates the effectiveness of our proposed method for winter wheat classification and phenology detection.

Detection of rice phenology through time series analysis of ground-based spectral index data

Monitoring crop phenology is of great importance for vegetation classification, yield estimation, and irrigation and fertilization management. To test the ability of ground-based remote sensing in detecting major phenological dates of rice, canopy spectra were collected by two portable spectrometers. The Red-edge Chlorophyll Index (CIred edge) and Normalized Difference Vegetation Index (NDVI) time-series derived from ground-based spectrometers were employed to detect the main specific phenological dates. CIred edge was obtained from ASD FieldSpec Pro spectrometer, while NDVI was from ASD FieldSpec Pro spectrometer (referred to as NDVIASD) and GreenSeeker RT 100 (referred to as NDVIGS). The phenology detection method consists of two procedures: (i) smoothing the temporal CIred edge and NDVI data with the double logistic regression function to represent intra-annual vegetation dynamics, (ii) determining the phenological dates through extracting the maximum, minimum and zero-crossing points (FDmax, FDmin and FDzero) from the first derivative value of the smoothed NDVI and CIred edge temporal profiles. A comparison of remote sensing-based estimates with field observations over three growing seasons with different cultivars, planting densities and nitrogen (N) rates showed that CIred edge can accurately estimate the dates of jointing, middle booting and dough grain. NDVI from both spectrometers can be used to detect the dates of active tillering, middle heading and maturity. Specifically, NDVIGS yielded better performance than NDVIASD for estimating the three phenological dates. Compared with growing season and planting density, rice cultivar and N rate exhibited more significant impact on the accuracy for phenology detection. This work has great potential to provide valuable support for assessing crop growth status and providing precise management strategy. The dates of active tillering, jointing and maturity detected from a combination of CIred edge and NDVI could be useful for irrigation and fertilization management, and harvest determination, respectively.

A hybrid approach for detecting corn and soybean phenology with time-series MODIS data

Abstract Monitoring crop phenology provides essential information for crop management, as well as for understanding regional to global scale vegetation dynamics. In this study, a hybrid phenology detection method is presented that incorporates the “shape-model fitting” concept of the two-step filtering method and a simulation concept of the crop models to detect the critical vegetative stages and reproductive stages of corn (Zea mays L.) and soybeans (Glycine max L.) from MODIS 250-m Wide Dynamic Range Vegetation Index (WDRVI) time-series data and 1000-m Land Surface Temperature (LST) data. The method was first developed and tested at the field scale over a ten-year period (2003–2012) for three experimental study sites in eastern Nebraska of USA, where the estimated phenology dates were compared to the ground-based phenology observations for both corn and soybeans. The average root mean square error (RMSE) of phenology stage estimation of the individual development stages across all sites ranged from 1.9 to 4.3 days for corn and from 1.9 to 4.9 days for soybeans. The approach was then tested at a regional scale over eastern Nebraska and the state of Iowa to evaluate its ability to characterize the spatio-temporal variation of targeted corn and soybean phenology stage dates over a larger area. Quantitative regional assessments were conducted by comparing the estimated crop stage dates with crop developmental stage statistics reported by the USDA NASS Crop Progress Reports (NASS-CPR) for both eastern Nebraska and Iowa. The accuracy of the regional-scale phenology estimation in Iowa (RMSE ranged from 2.6 to 3.9 days for corn and from 3.2 to 3.9 days for soybeans) was slightly lower than in eastern Nebraska (RMSE ranged from 1.8 to 2.9 days for corn and from 1.7 to 2.9 days for soybeans), However, the estimation accuracy in Iowa was still reasonable with the estimated phenology dates being within 4 days or less of the observed dates for both corn and soybeans.

Estimation of winter wheat phenology under the influence of cumulative temperature and soil salinity in the Yellow River Delta, China, using MODIS time-series data

ABSTRACTUnderstanding the impact of environmental factors on crop phenology is significant in predicting crop growth stages, agricultural decision-making, and yield estimation. Here, using Moderate Resolution Imaging Spectroradiometer time-series data, we present phenological detection mechanisms and an explanation for the phenological variability linked to environmental drivers, such as cumulative temperature and soil salinity, for winter wheat (Triticum aestivum L.) in the Yellow River Delta in 2013. The 8-day normalized difference vegetation index was fitted to a double Gaussian function. Phenological phases, such as the green-up and heading phases, were extracted using maximum curvature approaches. The spatial characteristics of the phenological patterns were investigated. The relationships between the phenological phases and cumulative temperature were explored. Then, the relationships between the phenological phases and soil salinity were evaluated by selecting sites with similar soil fertility and temperature forcing. This study concluded that the regional average green-up date occurred on 5 March, and the regional average heading date occurred on 9 May. The spatial distributions of the green-up and heading phases showed a gradual delay from the southwest to the northeast and from the south to the north. The green-up phase lagged 4–5 days for every 10 degree days that the cumulative temperature decreased. The heading phase lagged 1–2 days for every 10 degree days that the cumulative temperature decreased. The green-up phase in a non-salinization region might be approximately 5–9 days earlier than that in a severe or moderate salinization region. The heading phase in a severe region might occur approximately 1–8 days earlier than that in a non-salinization or moderate salinization region. The method proposed in this article may be useful for understanding the impact of temperature and soil salinity on phenology and could be used to better manage winter wheat in coastal salinization areas.

Selecting the Optimal NDVI Time-Series Reconstruction Technique for Crop Phenology Detection

AbstractA new scored method has been proposed in this study to evaluate the performances of different NDVI time-series reconstruction techniques. By giving a synthetic score to each of the candidates techniques based on two quantified criteria the optimal one is selected for the purpose of phenology detection. Three widely used techniques including Asymmetric Gaussian function fitting (AG), Double Logistic function fitting (DL) and Savitzky-Golay filtering (SG) are compared using NDVI time-series products from Moderate Resolution Imaging Spectroradiometer (MODIS) on Terra satellite over cropland of Northeast China. The results show that AG approach outperforms the two others in our study area. Cropland NDVI values have been improved obviously after the reconstruction by AG. Spatial patterns of the crop phenology detected from the AG reconstructed NDVI time-series are reasonable. The errors of the derived crop phenology metrics are within an acceptable limit.

Codominant water control on global interannual variability and trends in land surface phenology and greenness.

Identifying the relative importance of climatic and other environmental controls on the interannual variability and trends in global land surface phenology and greenness is challenging. Firstly, quantifications of land surface phenology and greenness dynamics are impaired by differences between satellite data sets and phenology detection methods. Secondly, dynamic global vegetation models (DGVMs) that can be used to diagnose controls still reveal structural limitations and contrasting sensitivities to environmental drivers. Thus, we assessed the performance of a new developed phenology module within the LPJmL (Lund-Potsdam-Jena managed Lands) DGVM with a comprehensive ensemble of three satellite data sets of vegetation greenness and ten phenology detection methods, thereby thoroughly accounting for observational uncertainties. The improved and tested model allows us quantifying the relative importance of environmental controls on interannual variability and trends of land surface phenology and greenness at regional and global scales. We found that start of growing season interannual variability and trends are in addition to cold temperature mainly controlled by incoming radiation and water availability in temperate and boreal forests. Warming-induced prolongations of the growing season in high latitudes are dampened by a limited availability of light. For peak greenness, interannual variability and trends are dominantly controlled by water availability and land-use and land-cover change (LULCC) in all regions. Stronger greening trends in boreal forests of Siberia than in North America are associated with a stronger increase in water availability from melting permafrost soils. Our findings emphasize that in addition to cold temperatures, water availability is a codominant control for start of growing season and peak greenness trends at the global scale.

Comparison of three NDVI time-series fitting methods in crop phenology detection in Northeast China

Phenological changes of cropland are the pivotal basis for farm management, agricultural production, and climate change research. Over the past decades, a range of methods have been used to extract phenological events based on satellite derived continuous vegetation index time series, however, large uncertainties still exist. In this study, three smoothing methods were compared to reduce the potential uncertainty and to quantify crop green-up dates over Northeast China. The results indicated that the crop spring onset dates estimated by three methods show variance in the dates, but with similar spatial pattern. In 60% of the study area, the standard deviation (SD) of the estimated starting date from different method is less than 10 days, while 39.5% of total pixels have SDs between 10days and 30 days. Through comparative analysis against the observation phenological data, we concluded that Asymmetric Gaussians produced the most approximative results of all, followed by Double Logistic algorithm, and Savizky-Glolay algorithm performed worst. The starting dates of crops occur mostly between May and June in this region. The Savitzky-Golay has the earliest estimates, while the Asymmetric Gaussians and Double logistic fitting method show similar and later estimates, which are more consistent with the observed data.

Remote Sensing Based Detection of Crop Phenology for Agricultural Zones in China Using a New Threshold Method

In recent years, the use of high temporal resolution satellite data has been emerging as an important tool to study crop phenology. Most methods to detect phenological events based on satellite data use thresholds to identify key events in the lifecycle of the crop. In this study, a new method was used to define such thresholds for identifying the start and end of the growing season (SOS/EOS) for 43 different agricultural zones in China. The method used 2000–2003 NOAA Advanced Very High Resolution Radiometer (AVHRR) satellite data with a spatial resolution of eight kilometers and a temporal resolution of 15 days. Following data pre-processing, time series for the normalized difference vegetation index (NDVI or N), slope of the NDVI curve (S), and difference (D) between the NDVI value and a base NDVI value for bare land without snow were constructed. For each zone, an optimal set of threshold values for N, D, and S was determined, based on the remote sensing data and observed SOS/EOS data for 2003 at 261 agro-meteorological stations. Results were verified by comparing the accuracy of the new proposed NDS threshold method with the results of three other methods for SOS/EOS detection with remote sensing data. The findings of all four methods were compared to in situ SOS/EOS data from 2000 to 2002 for 110 agro-meteorological stations. Results show that the developed NDS threshold method had a significantly higher accuracy compared with other methods. The method is mainly limited by the observed data and the necessity of reestablishing the thresholds periodically.

Determination of the gas exchange phenology in an evergreen coniferous forest from 7 years of eddy covariance flux data using an extended big‐leaf analysis

AbstractWe defined gas exchange phenology as the seasonality of the gas exchange characteristics of a forest canopy, and investigated how the gas exchange phenology could be directly detected from an eddy covariance (EC) dataset and its influence on the canopy fluxes within an evergreen Japanese cypress forest. For the detection of gas exchange phenology, we derived three bulk parameters of the extended big‐leaf model (Kosugi et al. 2005) inversely from EC flux data over a 7‐year period: surface conductance (gc), maximum rate of carboxylation of the “big leaf” (VCMAX), and intercellular CO2 concentration of the “big leaf” (CI). The relationship between gc and the vapor pressure deficit declined in winter and spring. The relationship between the daily ecosystem respiration and air temperature was greater in the spring than in the other seasons. The temperature dependence curve of VCMAX decreased substantially in the winter and was different from that of an evergreen broadleaved forest. A decrease in CI was occasionally coupled with the decrease in canopy gross primary production during April and August, indicating that stomatal closure was responsible for a decline in canopy photosynthesis. Gas exchange phenology should be quantified when understanding the determining factors of the seasonality of canopy fluxes at evergreen coniferous forests.

MODIS-based corn grain yield estimation model incorporating crop phenology information

A crop yield estimation model using time-series MODIS WDRVI was developed. The main feature of the proposed model is the incorporation of crop phenology detection using MODIS data, called the “Shape-Model Fitting Method”. MODIS WDRVI taken 7–10days before the corn silking stage had strong linear correlation with corn final grain yield at both field and regional scales. The model revealed spatial patterns of corn final grain yield all over the U.S. from 2000 to 2011. State-level corn yield was estimated accurately with coefficient of variation below 10% especially for the 18 major corn producing states including Iowa, Illinois, Delaware, Minnesota, Ohio, West Virginia, Wisconsin, Michigan, Indiana, Nebraska, Kentucky, New York, South Dakota, Missouri, Pennsylvania, Tennessee, New Jersey and Maryland. The results corresponded well with the spatial pattern of high-yield regions derived from the USDA/NASS data. However, the model tended to underestimate corn grain yield in three irrigated regions: the Midwestern region depending on the Ogallala Aquifer, the downstream basin of the Mississippi, and the southwestern region of Georgia. In contrast, it tended to overestimate corn grain yield around the outlying regions of the U.S. Corn Belt, specifically, the East Coast, North Dakota, Minnesota, Wisconsin, and Missouri. The estimation accuracy of the proposed model differed depending on the region. However, the annual variation of state level corn grain yield could be detected with high accuracy, especially in the major corn producing states.

Estimation of Water Consumption of Lowland Rice in Tropical Area based on Heterogeneous Cropping Calendar Using Remote Sensing Technology

Abstract Irrigated agricultural sector meets challenge in using water efficiently and effectively while improving food acquisition. One of the strategies to increase food production dealing with water scarcity is by increasing crop water productivity. Spatial data of crop water productivity parameters on large area is needed to support better water resources management. Remote sensing provides viable means to meet requirement of improved regional-scale data set of crop water productivity, such as crop water consumption (evapotranspiration) and crop yield. A methodology is developed to quantify spatial variation of seasonal evapotranspiration (ET) of lowland rice in paddy field with heterogeneous cropping calendar using satellite data. The methodology is developed by synthesizing SEBAL (Surface Energy Balance Algorithm for Land) method and lowland rice phenology detection method. Study area is located in the northern part of West Java Province, known as Pantura Jawa Barat region. The developed method is able to describe spatial variability of lowland rice evapotranspiration in Pantura Jawa Barat region for the growing period of dry season year 2004. The average seasonal actual evapotranspiration of rice is found to be 274.5 mm/season with standard deviation ± 44.7 mm/season.

Evaluation of Sub-Pixel Cloud Noises on MODIS Daily Spectral Indices Based on in situ Measurements

Cloud contamination is one of the severest problems for the time-series analysis of optical remote sensing data such as vegetation phenology detection. Sub-pixel clouds are especially difficult to identify and remove. It is important for accuracy improvement in various terrestrial remote sensing applications to clarify the influence of these residual clouds on spectral vegetation indices. This study investigated the noises caused by residual sub-pixel clouds on several frequently-used spectral indices (NDVI, EVI, EVI2, NDWI, and NDII) by using in situ spectral data and sky photographs at the satellite overpass time. We conducted in situ continuous observation at a Japanese deciduous forest for over a year and compared the MODIS spectral indices with the cloud-free in situ spectral indices. Our results revealed that residual sub-pixel clouds potentially contaminated about 40% of the MODIS data after cloud screening by the state flag of MOD09 product. These residual clouds significantly decreased NDVI values during the leaf growing season. However, such noises did not appear in the other indices. This result was thought to be caused by the different combination of wavelengths among spectral indices. Our results suggested that the noises by residual sub-pixel clouds can be reduced by using EVI, NDWI, or NDII in place of NDVI.

Detecting Spatiotemporal Changes of Corn Developmental Stages in the U.S. Corn Belt Using MODIS WDRVI Data

The dates of crop developmental stages are important variables for many applications including assessment of the impact of abnormal weather on crop yield. Time-series 250-m vegetation-index (VI) data acquired from the Moderate Resolution Imaging Spectroradiometer (MODIS) provide valuable information for monitoring the spatiotemporal changes of corn growth across large geographic areas. The goal of this study is to evaluate the performance of a new crop phenology detection method, namely, two-step filtering (TSF), for revealing the spatiotemporal pattern of specific corn developmental stages (early vegetative: V2.5; silking: R1; dent: R5; mature: R6) over an eight-year period (2001-2008) across Iowa, Illinois, and Indiana using MODIS derived Wide Dynamic Range VI data. Weekly crop progress reports produced by the U.S. Department of Agriculture National Agricultural Statistics Service (NASS) were used to assess the accuracy of TSF-based estimates of corn developmental stages. The results showed that the corn developmental stages could be estimated with high accuracy (the root mean squared error ranged from 4.1 to 5.5 days, the determination coefficient ranged from 0.66 to 0.84, and the coefficient of variation ranged from 2.1% to 3.7%) based on NASS-derived statistics on an agricultural statistics district level. In particular, the annual changes in the spatiotemporal patterns of the estimated silking stage had a high level of agreement with those of the NASS-derived statistics. These results suggested that the TSF method could provide local-scale information of corn phenological stages, which had an advantage over the NASS-derived statistics particularly in terms of the spatial resolution.

Evaluation of compositing period and AVHRR and MERIS combination for improvement of spring phenology detection in deciduous forests

Accurate and precise detection of phenology events is needed to assess trends in seasonal vegetation development indicative of climate or other environmental change processes. In this research, detection accuracy of start of season (SOS) phenology for deciduous forest across Eastern Canada was assessed using satellite time series and in situ PlantWatch observations. Several aspects were evaluated regarding performance of phenology information extraction: 1) effect of compositing period, 2) individual performance of the Advanced Very High Resolution Radiometer (AVHRR) and the Medium Resolution Imaging Spectrometer (MERIS) sensors, and 3) performance for these sensors combined. The AVHRR and MERIS sensors were used as they are overlapping operational missions with planned future continuity. Three approaches to utilizing the multi-sensor data were tested: 1) inter-calibrating NDVI data between sensors and using the multi-sensor data stream to detect SOS, 2) combining independently derived SOS estimates from AVHRR and MERIS based on a weighted average, and 3) combining approaches 1 and 2. Comparison with in situ observations of leaf out and first bloom showed that combining independent SOS estimates from AVHRR and MERIS was better than using the inter-calibrated multi-sensor data. Combining SOS estimates from both sensors reduced error by 1–2 days compared to the individual sensor results. Composite periods from 7 to 11 days produced the best results for leaf out with a mean absolute error (MAE) of 5 days. Results for first bloom were not as good as those for leaf out, producing a MAE of 6.5 days. For first bloom, compositing periods greater than 11 days did not increase error at the same rate as seen for leaf out. However, the larger MAE observed for first bloom may have masked this effect.

Detecting Rice Phenology in Paddy Fields with Complex Cropping Pattern Using Time Series MODIS Data A Case Study of Northern West Java – Indonesia

Monitoring paddy rice phenology and cropping schedules over wide areas is essential for many applications. Remote sensing provides a viable means to meet the requirement of improved regional-scale data set of paddy rice fields, such as phenological stages. A number of methods have been developed for detecting seasonal vegetation changes by using satellite images. Development of such methods to paddy fields with complex cropping pattern is still challenging. In this study, we developed a method for remotely determining phenological stages of paddy rice that uses time series of two vegetation indices (EVI and LSWI) obtained from MODIS data. We ran the algorithm to determine the planting date, heading date, and harvesting date of paddy rice in 5 districts of West Java Province, using the 8-day composite MODIS Surface Reflectance products (500-m spatial resolution) in 2004. Estimated harvesting dates were then used to calculate paddy rice harvested area. We validated the performance of the method against statistical data in 13 subdistricts. The root mean square errors of the estimated paddy rice harvested area against the statistical data were: 851 Ha for monthly data, 1227 Ha for quarterly data, and 2433 Ha for yearly data. Subdistrict-level comparisons of paddy rice harvested area between the MODIS estimation and statistical data showed moderate correlation, with coefficient of determination (r2) 0.6, 0.7, and 0.6 for monthly, quarterly and yearly data, respectively. The results of this study indicated that the MODIS-based paddy rice phenological detection algorithm could potentially be applied at large spatial scales to monitor paddy rice agriculture on a timely and frequent basis.