How Well Does the ‘Small Fire Boost’ Methodology Used within the GFED4.1s Fire Emissions Database Represent the Timing, Location and Magnitude of Agricultural Burning?
"> Figure 1
<p>Spatial mapping (0.25°) of the fire affected regions for 2015 in eastern China in June (<b>a</b>–<b>d</b>), August (<b>e</b>–<b>h</b>) and October (<b>i</b>–<b>l</b>). (1st column) burned area (BA) reported by GFED4, which uses the MODIS MCD64A1 BA product as its BA metric for all years after the early 2000’s. (2nd column) BA reported by the most recent GFED4.1s inventory, which increases the basic MCD64A1 measured BA via an adaptation of the ‘small fire boosting’ approach developed by [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>] (which is primarily driven by MODIS AF detections). (3rd column) Fraction of BA reported by GFED4.1s that comes from the SFB methodology of [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>], and (4th column) the fire radiative power (FRP) reported in the same area by the VIIRS-IM product of Zhang et al. [<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>] that is sensitive to both small and large fires. June and October are the periods of most intensive agricultural burning in this region of eastern China [<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>,<a href="#B21-remotesensing-10-00823" class="html-bibr">21</a>], which makes the ‘strongly boosted’ BA seen in August in GFED4.1s (highlighted with the blue dashed rectangle in (<b>f</b>)) seem potentially erroneous, and at odds with a lack of fire activity reported by the VIIRS-IM product for the same month (<b>h</b>).</p> "> Figure 2
<p>Monthly burned area (0.25°) reported for 2016 across India by the MCD64A1 product, as derived from 500 m MODIS observations using the algorithm of [<a href="#B4-remotesensing-10-00823" class="html-bibr">4</a>] and reported by GFED4. In Punjab, north-western India (highlighted by the blue rectangle) fire shows a strong seasonal pattern, with activity occurring primarily in May and then again (stronger) in October and November. These periods are associated primarily with agricultural fires, with wheat stubble primarily burned in May and rice straw in October–November.</p> "> Figure 3
<p>MODIS 1 km active fire (AF) detections produced using the algorithm described in [<a href="#B4-remotesensing-10-00823" class="html-bibr">4</a>] and displayed in red within two 0.25 deg × 0.25 deg areas (representing the areas of two GFED grid cells) in eastern China, overlaying Sentinel-2 true colour composites (centered at 31.875° N,120.625° E and 31.625° N,120.875° E respectively, retrieved on 22 August 2016). Every one of the MODIS AF detections can be seen to be located in an area where outdoor vegetation burning is unlikely to have occurred, for example in residential and industrial areas. None are on agricultural land, and it is therefore very likely that these AF detections are false alarms caused by manmade structures being warmer or more glint-inducing than their surroundings—as detailed in References [<a href="#B7-remotesensing-10-00823" class="html-bibr">7</a>,<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>]. Such effects mean AF detection algorithms are frequently subject to these types of false alarms in eastern China, and Reference [<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>] developed methods to mask these false alarms based on temporal persistence and landcover information. The MCD14 AF product does not apply such masking, so the false alarms remain, and their presence in areas not matched by MODIS burned area detections causes GFED4.1s to report significant burned area for these two grid cells (24 and 19 km<sup>2</sup>, respectively). These large burned areas come from the application of the ‘boosting’ methodology of [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>] which is used within GFEDv4.1s to ‘boost’ burned area undetected by the MODIS MCD64A1 BA product. This SFB method takes MODIS AF detections located at sites where there is matching MCD64A1 detected burned area and, once a calibration between AF count and BA size is developed, uses the AF’s detected in areas where no MCD64A1 burned area detection occurs to estimate the ‘undetected’ BA from the presumed ‘small fires’ (see <a href="#remotesensing-10-00823-f001" class="html-fig">Figure 1</a>f for locations of these cells and [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>] for full details of the SFB methodology).</p> "> Figure 4
<p>MODIS active fire (AF) pixel counts from MCD14 (0.25° grid cell resolution) in the eastern China region highlighted by the blue dashed box in <a href="#remotesensing-10-00823-f001" class="html-fig">Figure 1</a>. AFs were classified via the landcover-based approach of Zhang et al. [<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>] as either suspected “true” active fire detections (left column), or false alarms (right column). The latter are identified by their occurrence in non-vegetated areas, such as urban regions (see [<a href="#B17-remotesensing-10-00823" class="html-bibr">17</a>] for more details). The ‘true’ and ‘false alarm’ AF counts are shown for June (upper row), August (middle row) and October 2015 (lower row), and false alarm AFs are clearly most apparent in August 2015 and in the region of eastern China highlighted in <a href="#remotesensing-10-00823-f001" class="html-fig">Figure 1</a>f as showing a potentially suspicious region of very significantly ‘boosted’ GFED4.1s burned area.</p> "> Figure 5
<p>Monthly fire activity for 2012–2015 for eastern China (area shown in <a href="#remotesensing-10-00823-f001" class="html-fig">Figure 1</a>). (<b>a</b>) GFED4-reported burned area (BA) (derived only from the MODIS MCD64A1 BA product), along with the additional BA reported in GFED4.1s that comes from the ‘small fire boost’ derived using the method reported in [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>]). Most of the GFED4.1s BA comes from the SFB rather than actual MODIS BA mapping. (<b>b</b>) ‘Boosted’ BA data classified as either BA derived from suspected true active fire detections, or BA derived from suspected AF false alarms (classifiedusing landcover-based masking; see <a href="#remotesensing-10-00823-f004" class="html-fig">Figure 4</a>).</p> "> Figure 6
<p>Total (<b>left column</b>) and cumulative time series (<b>right column</b>) of fire activity for October and November, 2016 for the 0.25° GFED grid cell located in the Punjab, north-western India (centred at 30.375° N, 74.875° E, shown spatially in <a href="#remotesensing-10-00823-f007" class="html-fig">Figure 7</a>). (<b>a</b>) comparison of total BA in October (blue) and November (red) 2016 reported by GFED4 (based on MCD64A1) and GFED4.1s (based on MCD64A1 + the SFB of [<a href="#B11-remotesensing-10-00823" class="html-bibr">11</a>]), along with that derived from Landsat and Sentinel-2 dNBR-based BA mapping, and (right column) daily cumulative (<b>b</b>) MCD64A1 BA, (<b>c</b>) MCD14 FRP, (<b>d</b>) VIIRS-IM FRP, and (<b>e</b>) dNBR-derived BA from Landsat/Sentinel-2.</p> "> Figure 7
<p>Detail of a 0.25° GFED grid cell centered at 30.375° N, 74.875° E, Punjab, India and imaged repeatedly over the October to November 2016 burning season. (<b>Left column</b>) true colour composites derived from Landsat/Sentinel-2 imagery, and (<b>right column</b>) <math display="inline"><semantics> <mrow> <mo>∆</mo> <mi>N</mi> <mi>B</mi> <mi>R</mi> </mrow> </semantics></math> map overlaid with burned area and AF detection data from MODIS and VIIRS collected during the interval between each Landsat/Sentinel-2 image. <math display="inline"><semantics> <mrow> <mo>∆</mo> <mi>N</mi> <mi>B</mi> <mi>R</mi> </mrow> </semantics></math> caption reports mean <math display="inline"><semantics> <mrow> <mo>∆</mo> <mi>N</mi> <mi>B</mi> <mi>R</mi> </mrow> </semantics></math> of the whole grid cell, and the amount of burned area present within the grid cell can be seen increasing from the 10 October 2016 to the 27 November 2016.</p> "> Figure 8
<p>Fire-relevant data covering a 0.25° GFED grid cell in (<b>top row</b>) eastern China (centred at 33.375° N, 114.625° E), and (<b>bottom row</b>) the Punjab (northwestern India, centred at 31.625° N, 75.375° E). (<b>a</b>–<b>c</b>) show Landsat NIR false colour composites taken respectively before (24 April 2015), during (11 June 2015) and after (19 June 2015) the fire activity in eastern China, and (<b>e</b>–<b>g</b>) before (24 September 2016), during (26 October 2016) and after (27 November 2016) the fire activity in the Punjab, India. For comparison, (<b>d</b>,<b>h</b>) show the mapped fire activity respectively reported in these same two areas during the same period by the 500 m MODIS MCD64A1 burned area (BA) products, the 1 km MCD14 active fire (AF) product, and the 375 m VIIRS-IM AF product. Compared to the evidence from Landsat, eastern China appears to have large areas of burning missing in the MCD64A1 BA product, whereas in the Punjab the MCD64A1 product appears to do relatively well south of the river that bisects the grid cell from southwest to northeast, but less so north of the river where the small agricultural fires appear more fragmented in the Landsat imagery.</p> "> Figure 9
<p>Fire activity within a 0.25° GFED grid cell centered at 33.625° N, 114.375° E in eastern China, May–June 2015? (<b>a</b>) Landsat-8 NIR false colour composite of 19 June 2015, after fire activity occurred in this grid cell. Recently burned areas appear dark in this rendition. (<b>b</b>) Spatial, and (<b>c</b>–<b>e</b>) temporal distribution of three fire products, MCD64A1 BA data, along with MCD14 and VIIRS-IM AF data. The spatial and temporal disagreement of MCD64A1 with MCD14, VIIRS-IM and the Landsat 8 imagery indicates that the burned area detected here by MCD64A1 seems likely to be caused by non-fire processes (e.g., crop harvesting) and is therefore represents an error of commission rather than true fire activity. When the fires do occur (as seen in the AF data) the MCD61A1 then fails to detect their ‘real’ burned area, resulting in a BA error of omission.</p> "> Figure 10
<p>Time series of fire datasets for eastern China (region shown in <a href="#remotesensing-10-00823-f001" class="html-fig">Figure 1</a>) during the summer burning season (May–June) from 2012–2015. The MODIS MCD64A1 product reports detected burned area (BA), and the MODIS MCD14DL and VIIRS-IM show retrieved fire radiative power (FRP) made at the locations of active fire (AF) detections. Colours represent different GFED grid cells within the region, and for clarity only those cells showing the fifty highest MCD64A1 burned area values are shown here. MCD64A1 reported burned area typically peaks 1 to 2 weeks before the AF products detect the peak in fire occurrence, suggesting that the BA product is identifying the initial harvesting of the crop rather than any subsequent burning of the crop residue left in the fields after harvest.</p> ">
Abstract
:1. Introduction
2. Datasets, Study Areas and Data Processing
2.1. Datasets
2.2. Study Areas
2.3. Landsat and Sentinel-2 Data Processing
3. Evaluation of GFED4.1s ‘Small Fire Boosting’
3.1. Eastern China
3.2. Punjab, India
4. Underlying Issues with MCD64A1 and MCD14
5. Summary and Conclusions
Author Contributions
Acknowledgments
Conflicts of Interest
References
- Seiler, W.; Crutzen, P.J. Estimates of gross and net fluxes of carbon between the biosphere and the atmosphere from biomass burning. Clim. Chang. 1980, 2, 207–247. [Google Scholar] [CrossRef]
- Van der Werf, G.R.; Randerson, J.T.; Giglio, L.; Collatz, G.J.; Mu, M.; Kasibhatla, P.S.; Morton, D.C.; DeFries, R.S.; Jin, Y.V.; van Leeuwen, T.T. Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997–2009). Atmos. Chem. Phys. 2010, 10, 11707–11735. [Google Scholar] [CrossRef] [Green Version]
- Giglio, L.; Randerson, J.T.; Werf, G.R. Analysis of daily, monthly, and annual burned area using the fourth-generation global fire emissions database (GFED4). J. Geophys. Res. Biogeosci. 2013, 118, 317–328. [Google Scholar] [CrossRef]
- Giglio, L.; Descloitres, J.; Justice, C.O.; Kaufman, Y.J. An enhanced contextual fire detection algorithm for MODIS. Remote Sens. Environ. 2003, 87, 273–282. [Google Scholar] [CrossRef]
- Giglio, L.; Van der Werf, G.R.; Randerson, J.T.; Collatz, G.J.; Kasibhatla, P. Global estimation of burned area using MODIS active fire observations. Atmos. Chem. Phys. 2006, 6, 957–974. [Google Scholar] [CrossRef] [Green Version]
- Giglio, L.; Loboda, T.; Roy, D.P.; Quayle, B.; Justice, C.O. An active-fire based burned area mapping algorithm for the MODIS sensor. Remote Sens. Environ. 2009, 113, 408–420. [Google Scholar] [CrossRef]
- Schroeder, W.; Oliva, P.; Giglio, L.; Csiszar, I.A. The New VIIRS 375 m active fire detection data product: Algorithm description and initial assessment. Remote Sens. Environ. 2014, 143, 85–96. [Google Scholar] [CrossRef]
- Wooster, M.J.; Roberts, G.; Perry, G.L.W.; Kaufman, Y.J. Retrieval of biomass combustion rates and totals from fire radiative power observations: FRP derivation and calibration relationships between biomass consumption and fire radiative energy release. J. Geophys. Res. Atmos. 2005, 110. [Google Scholar] [CrossRef]
- Roy, D.P.; Boschetti, L.; Justice, C.O.; Ju, J. The collection 5 MODIS burned area product—Global evaluation by comparison with the MODIS active fire product. Remote Sens. Environ. 2008, 112, 3690–3707. [Google Scholar] [CrossRef]
- Van der Werf, G.R.; Randerson, J.T.; Giglio, L.; van Leeuwen, T.T.; Chen, Y.; Rogers, B.M.; Mu, M.; van Marle, M.J.; Morton, D.C.; Collatz, G.J.; et al. Global fire emissions estimates during 1997–2016. Earth Syst. Sci. Data 2017, 9, 697. [Google Scholar] [CrossRef]
- Randerson, J.T.; Chen, Y.; Werf, G.R.; Rogers, B.M.; Morton, D.C. Global burned area and biomass burning emissions from small fires. J. Geophys. Res. Biogeosci. 2012, 117. [Google Scholar] [CrossRef] [Green Version]
- Chan, C.K.; Yao, X. Air pollution in mega cities in China. Atmos. Environ. 2008, 42, 1–42. [Google Scholar] [CrossRef]
- Chen, J.; Li, C.; Ristovski, Z.; Milic, A.; Gu, Y.; Islam, M.S.; Wang, S.; Hao, J.; Zhang, H.; He, C.; et al. A review of biomass burning: Emissions and impacts on air quality, health and climate in China. Sci. Total Environ. 2017, 579, 1000–1034. [Google Scholar] [CrossRef] [PubMed]
- Granier, C.; Bessagnet, B.; Bond, T.; D’Angiola, A.; van Der Gon, H.D.; Frost, G.J.; Heil, A.; Kaiser, J.W.; Kinne, S.; Klimont, Z.; et al. Evolution of anthropogenic and biomass burning emissions of air pollutants at global and regional scales during the 1980–2010 period. Clim. Chang. 2011, 109, 163. [Google Scholar] [CrossRef]
- Sharma, A.R.; Kharol, S.K.; Badarinath, K.V.S.; Singh, D. Impact of agriculture crop residue burning on atmospheric aerosol loading—a study over Punjab State, India. Ann. Geophys. 2010, 28, 367. [Google Scholar] [CrossRef]
- Zhang, T.; Wooster, M.J.; Green, D.C.; Main, B. New field-based agricultural biomass burning trace gas, PM2. 5, and black carbon emission ratios and factors measured in situ at crop residue fires in Eastern China. Atmos. Environ. 2015, 121, 22–34. [Google Scholar] [CrossRef]
- Zhang, T.; Wooster, M.J.; Xu, W. Approaches for synergistically exploiting VIIRS I-and M-Band data in regional active fire detection and FRP assessment: A demonstration with respect to agricultural residue burning in Eastern China. Remote Sens. Environ. 2017, 198, 407–424. [Google Scholar] [CrossRef]
- National Bureau of Statistics of China (NBSC). China Statistical Yearbook 2012; National Bureau of Statistics of China (NBSC): Beijing, China, 2012.
- Streets, D.G.; Bond, T.C.; Carmichael, G.R.; Fernandes, S.D.; Fu, Q.; He, D.; Klimont, Z.; Nelson, S.M.; Tsai, N.Y.; Wang, M.Q.; et al. An inventory of gaseous and primary aerosol emissions in Asia in the year 2000. J. Geophys. Res. Atmos. 2003, 108. [Google Scholar] [CrossRef]
- Vadrevu, K.P.; Ellicott, E.; Badarinath, K.V.S.; Vermote, E. MODIS derived fire characteristics and aerosol optical depth variations during the agricultural residue burning season, north India. Environ. Pollut. 2011, 159, 1560–1569. [Google Scholar] [CrossRef] [PubMed]
- Jain, N.; Bhatia, A.; Pathak, H. Emission of air pollutants from crop residue burning in India. Aerosol Air Qual. Res. 2014, 14, 422–430. [Google Scholar] [CrossRef]
- French, N.H.F.; Kasischke, E.S.; Hall, R.J.; Murphy, K.A.; Verbyla, D.L.; Hoy, E.E.; Allen, J.L. Using Landsat data to assess fire and burn severity in North American boreal forest region: An overview and summary of results. Int. J. Wildl. Fire 2008, 17, 443–462. [Google Scholar] [CrossRef]
- Mandanici, E.; Bitelli, G. Preliminary Comparison of Sentinel-2 and Landsat 8 Imagery for a Combined Use. Remote Sens. 2016, 8, 1014. [Google Scholar] [CrossRef]
- Huang, X.; Li, M.; Li, J.; Song, Y. A high-resolution emission inventory of crop burning in fields in China based on MODIS Thermal Anomalies/Fire products. Atmos. Environ. 2012, 50, 9–15. [Google Scholar] [CrossRef]
- Giglio, L.; Boschetti, L.; Roy, D.; Hoffmann, A.A.; Humber, M. Collection 6 MODIS Burned Area Product User’s Guide Version 1.0. 2016. Available online: http://modis-fire.umd.edu/pages/manuals.php (accessed on 20 May 2018).
Dataset | System | Resolution |
---|---|---|
GFED4 Burned Area | MODIS, ATSR, VIIRS | 0.25° |
GFED4.1s Burned Area | ||
MCD64A1 Burned Area | MODIS | 500 m |
MCD14 Active Fire | MODIS | 1 km |
VIIRS Regional Active Fire | VIIRS | 375 m |
Landsat Level 1B product | Landsat 7/8 | 30 m |
Sentinel-2 Level 1B product | Sentinel-2 MSI | 10, 20, 60 m |
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhang, T.; Wooster, M.J.; De Jong, M.C.; Xu, W. How Well Does the ‘Small Fire Boost’ Methodology Used within the GFED4.1s Fire Emissions Database Represent the Timing, Location and Magnitude of Agricultural Burning? Remote Sens. 2018, 10, 823. https://doi.org/10.3390/rs10060823
Zhang T, Wooster MJ, De Jong MC, Xu W. How Well Does the ‘Small Fire Boost’ Methodology Used within the GFED4.1s Fire Emissions Database Represent the Timing, Location and Magnitude of Agricultural Burning? Remote Sensing. 2018; 10(6):823. https://doi.org/10.3390/rs10060823
Chicago/Turabian StyleZhang, Tianran, Martin J. Wooster, Mark C. De Jong, and Weidong Xu. 2018. "How Well Does the ‘Small Fire Boost’ Methodology Used within the GFED4.1s Fire Emissions Database Represent the Timing, Location and Magnitude of Agricultural Burning?" Remote Sensing 10, no. 6: 823. https://doi.org/10.3390/rs10060823